Upload
phunghanh
View
223
Download
2
Embed Size (px)
Citation preview
General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Jun 14, 2018
2nd Generation Alkaline ElectrolysisFinal report
Yde, Lars; Kjartansdóttir, Cecilía Kristín; Allebrod, Frank; Mogensen, Mogens Bjerg; Møller, Per; Hilbert,Lisbeth R.; Nielsen, Peter Tommy; Mathiesen, Troels; Jensen, Jørgen; Andersen, Lars; Dierking,Alexander
Publication date:2013
Document VersionPublisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):Yde, L., Kjartansdóttir, C. K., Allebrod, F., Mogensen, M. B., Møller, P., Hilbert, L. R., ... Dierking, A. (2013). 2ndGeneration Alkaline Electrolysis: Final report. Århus University Business and Social Science – Centre for EnergyTechnologies.
1
2nd Generation Alkaline Electrolysis
FORCE Technology
GreenHydrogen.dk.
Technical University of Denmark – Energy Conversion
Technical University of Denmark – Mechanical Engineering
Århus University Business and Social Science – Centre for Energy Technologies (CET (former HIRC))
Final report
EUDP 63011-0200
March 2013
2
Content Content .............................................................................................................................................................. 2
1. Introduction ............................................................................................................................................... 4
2. Project Summary and Conclusions ............................................................................................................ 5
3. Background and Project Objective .......................................................................................................... 10
4. Structure of the report ............................................................................................................................ 12
4.1. Improved electrodes (WP1)............................................................................................................. 12
4.2. Higher operation temperature (WP2) ............................................................................................. 12
4.3. Higher operation pressure (WP3) .................................................................................................... 12
4.4. Improved electrolysis stacks architecture (WP4) ............................................................................ 13
4.5. Corrosion (WP5) .............................................................................................................................. 13
4.6. Modular design for easy up scaling (WP6) ...................................................................................... 13
4.7. Demonstration (WP7) ...................................................................................................................... 13
5. Improved electrodes (WP 1) .................................................................................................................... 14
5.1. Track 1 Electrochemical Plating ....................................................................................................... 14
5.1.1. Literature study and preliminary experiments ........................................................................ 14
5.1.2. Electrode development - Processes for producing a stable and durable porous nickel
electrode 18
5.1.3. Electrochemical measurements .............................................................................................. 21
5.1.4. Durability test .......................................................................................................................... 23
5.1.5. Summary .................................................................................................................................. 24
5.1.6. References ............................................................................................................................... 24
5.2. Track 2 Atmospheric Plasma Technology ........................................................................................ 26
5.2.1. Development of improved electrodes (WP 1.1-1.6) ............................................................... 26
5.2.2. Manufacturing of electrodes by plasma spraying (WP 1.7) .................................................... 29
5.2.3. Long-term tests and storage of plasma-sprayed electrodes (WP 1.8) .................................... 29
5.2.4. References ............................................................................................................................... 29
5.3. Track 3 HTP Materials ...................................................................................................................... 30
5.3.1. Abstract of Frank Allebrod Ph.D. thesis “High temperature and pressure alkaline electrolysis”
30
5.3.2. Important literature data ........................................................................................................ 31
5.3.3. Constructed measurement set-ups ......................................................................................... 35
5.3.4. Specific electrical conductivity measurements on free and immobilized aqueous KOH ........ 41
3
5.3.5. Alkaline electrolysis cell for 250 C and 40 bar ....................................................................... 42
5.3.6. “Long term” galvanostatic test ................................................................................................ 52
5.3.7. Summary .................................................................................................................................. 53
5.3.8. List of publications of the project from DTU Energy Conversion ............................................ 54
5.4. Results on improved electrodes ...................................................................................................... 54
5.4.1. Electroplating ........................................................................................................................... 54
5.4.2. Atmospheric Plasma Spray ...................................................................................................... 55
5.4.3. HTP electrodes ......................................................................................................................... 55
5.4.4. Conclusion ............................................................................................................................... 56
6. Higher operation pressure (WP 3) ........................................................................................................... 57
7. Improved electrolysis stack architecture (WP 4) ..................................................................................... 59
8. Corrosion resistant materials (WP 5) ...................................................................................................... 61
9. Modular design for easy up scaling (WP 6) ............................................................................................. 63
9.1. Electrolyser module ......................................................................................................................... 63
9.2. Deoxer Module ................................................................................................................................ 64
9.3. Dryer Module................................................................................................................................... 65
9.4. Water Treatment Module ............................................................................................................... 65
9.5. Power supply and Control unit ........................................................................................................ 66
9.6. Rack Mounted Electrolyser .............................................................................................................. 66
10. Demonstration (WP7) .......................................................................................................................... 67
10.1. Building and Installation .............................................................................................................. 67
10.2. Measurement Results .................................................................................................................. 68
10.2.1. Stack efficiency ........................................................................................................................ 68
10.2.2. System Efficiency ..................................................................................................................... 68
10.2.3. Total production hours ............................................................................................................ 68
10.2.4. Gas purity ................................................................................................................................. 68
10.2.5. Cooling ..................................................................................................................................... 69
10.3. Demonstration ............................................................................................................................. 69
11. List of publications ............................................................................................................................... 70
4
1. Introduction This report provides the results of the 2nd Generation Alkaline Electrolysis project which was initiated in
2008. The project has been conducted from 2009-2012 by a consortium comprising Århus University
Business and Social Science – Centre for Energy Technologies (CET (former HIRC)), Technical University of
Denmark – Mechanical Engineering (DTU-ME), Technical University of Denmark – Energy Conversion (DTU-
EC), FORCE Technology and GreenHydrogen.dk. The project has been supported by EUDP.
Authors and contributors to the project:
Lars Yde, CET
Cecilía K. Kjartansdóttir, DTU-ME
Frank Allebrod, DTU-EC
Mogens B. Mogensen, DTU-EC
Per Møller, DTU-ME
Lisbeth R. Hilbert, FORCE Technology
Peter Tommy Nielsen, FORCE Technology
Troels Mathiesen, FORCE Technology
Jørgen Jensen, GreenHydrogen.dk
Lars Andersen, GreenHydrogen.dk
Alexander Dierking, GreenHydrogen.dk
Edited by Anne Nielsen, CET
The overall project administration has been conducted by CET.
5
2. Project Summary and Conclusions In order to keep production of renewable energy in balance with the consumption load management is
needed. Production of hydrogen by water electrolysis can serve as a huge balancing load and at the same
time act as an energy storage media.
The overall purpose of this project has been to contribute to this load management by developing a 2nd
generation of alkaline electrolysis system characterized by being compact, reliable, inexpensive and energy
efficient. The specific targets for the project have been to:
Increase cell efficiency to more than 88% (according to the higher heating value (HHV)) at a
current density of 200 mA /cm2
Increase operation temperature to more than 100 degree Celsius to make the cooling energy
more valuable
Obtain an operation pressure more than 30 bar hereby minimizing the need for further
compression of hydrogen for storage
Improve stack architecture decreasing the price of the stack with at least 50%
Develop a modular design making it easy to customize plants in the size from 20 to 200 kW
Demonstrating a 20 kW 2nd generation stack in H2College at the campus of Århus University in
Herning
These targets are well in line with the objectives which were established in the 2009 strategy for RD&D on
electrolysis in Denmark for the period 2011-13. The targets for alkaline electrolysis stacks can be
summarised as follows:
Energy efficiency 88 %
Cell voltage: 1,68 V
Current density: 200 mA/cm2
Operation temperature: 100 °C
Operation pressure: 30 bar
In order to achieve the targets the project has been organised in the following work packages, which
roughly creates the structure of this report:
WP1: Improved electrodes
WP2: Higher operation temperature
WP3: Higher operation pressure
WP4: Improved electrolysis stacks architecture
WP5: Corrosion
WP6: Modular design for easy up scaling
WP7: Demonstration
6
Improved electrodes and elevated operation temperatures (WP1 and 2)
The project has included research and development on three different technology tracks of electrodes; an
electrochemical plating, an atmospheric plasma spray (APS) and finally a high temperature and pressure
(HTP) track with operating temperature around 250 C and pressure around 40 bar. The HTP track was
misleadingly named ‘the ceramic track’ in the initial project description submitted with the EUDP
application. The more correct term ‘high temperature and pressure track’ (HTP) will be used for the rest of
this report. Both the electrochemical plating and the HTP track have each involved a PhD study, which is
reflected in this report.
The results obtained in WP 1 show that all three electrode tracks have reached high energy efficiencies. In
the electrochemical plating track a stack efficiency of 86.5% at a current density of 177mA/cm2 and a
temperature of 74.4 ˚C has been shown. The APS track showed cell efficiencies of 97%, however, coatings
for the anode side still need to be developed. The HTP cell has reached 100 % electric efficiency operating
at 1.5 V (the thermoneutral voltage) with a current density of 1. 1 A/cm2. This track only tested small cells
in an externally heated laboratory set-up, and thus the thermal loss to surroundings cannot be given.
Elevation of operation temperature has been a part of the PhD study on HTP electrodes. Literature studies
as well as experimental work have shown that elevation of operating temperature causes remarkably
higher current density entailing higher cell efficiency.
As the HTP track is still on research level and the elevated temperature combined with the highly corrosive
electrolyte KOH require highly resistant materials for the electrodes the elevated temperatures have not
been pursued for the other two tracks.
Based on the findings in the project GreenHydrogen.dk has assessed that the electroplating track on the
short term gives the most promising performance when combining durability, lifetime, efficiency and
production price. This track has been chosen for further development and demonstration in three projects;
’Optimisation of 2nd Generation Alkaline Elektrolysis’, ’Hyprovide’ and ’High-efficiency, low-cost electrode-
surfaces for next generation alkaline electrolysis’, which respectively pursue optimisation of the production
of electrodes and scaling up the electrolysis system to MW size.
Higher operation pressure (WP 3)
The goal set for the 2nd generation electrolyser system, has been to generate 30 bar pressure in the cell
stack. Water electrolysis generates hydrogen at a very high pressure, which provides the possibility to
generate compressed hydrogen directly from the cell stack, without spending additional power on
compression afterwards.
Further, high pressure electrolysis offer the advantage of reduced size of tubing and components, which
provides a possibility to increase the capacity or reduce the size of the components by a factor equal to the
pressure ratio. This enables the electrolyser system to be designed with a significantly smaller footprint
than equivalent atmospheric systems.
7
An obstacle to be overcomed has been to ensure equalisation of the H2 and O2 pressure to avoid that
mixing of gasses can occur. To solve this problem, a special equilibrium valve has been developed to
mechanically control that the pressure of the H2 at all times equals the O2 side.
The developments in this work package have resulted in a stack design, which is a cylindrical pressure
vessel, with each cell having a cell “wall” sufficiently thick, to resist the high pressure and sealed with O-
rings for perfect sealing at high pressures. The stack has in test proved to resist a pressure on 45 bar,
though some adjustment is still needed to optimize the pressure resistance and efficiency.
Improved electrolysis stack architecture (WP 4)
The primary goal of WP4 has been to develop an electrolyser stack concept suitable for mass production,
while still solving the technical challenges when designing a high pressure alkaline electrolyser stack. The
challenges are mechanical strength, electrolyte and gas flow, stray currents and galvanic corrosion.
When deciding on the new stack design both a ‘zero gap’ and ‘non-zero gap’ was considered. The zero gap
design is more efficient than non-zero gap, however the design is more complex and very costly, primarily
because the additional materials and production costs for zero gap electrodes..
From these considerations, the concept of a “low gap”, low diameter, high pressure and high cell number
electrolyser stack was born, which could offer an improved efficiency of the electrolyser without causing
the same high material and production cost as a zero gap zero gap solution.
As a result the low gap design and pressurized stack has reduced the price by 60% of the total system, as
well as a reduced system footprint.
Corrosion resistant materials (WP 5)
Hydrogen production at increased temperature, pressure and stack voltage requires that the electrolyser
can withstand the aggressive alkaline media under those severe conditions. A design life of more than 10
years is a challenge in terms of materials.
The progress of the project required a special focus on corrosion testing and examination of polymers in
order to find alternative durable membrane and gasket materials.
The initial literature survey and the first tests at H2College indicated that the chemical resistance of
polymers presented a greater challenge than anticipated, and that test data from commercial suppliers
were insufficient to model the conditions in the electrolyser. The alkali resistant polymers (e.g. Teflon) are
costly and the search for cheaper alternatives turned into a major aim.
A number of different tests were run under accelerated conditions and the degradation mechanism was
examined. The results of standardised tests are summarised in Table 5 on page 61. Consequently, the
project has involved corrosion testing and examination of polymers to greater extent than originally
planned at the detriment of long-term testing of electrodes (WP 1.8), which has been limited.
8
Inspiration for the future materials selection in electrolysers is now available by the combined experience
from the experimental work and literature.
Modular design for easy up scaling (WP 6)
One of the strong visions for the 2nd generation electrolyser concept has been to develop a modular system
with several standardized components, making it possible to deliver a wide range of specification and
features to the customer depending on the application. The use of a modular system is also aimed at faster
and easier service, by replacing defect modules directly on site to decrease the downtime.
A great effort has been put into analysing the electrolysis system to determine how the different
components could be split onto modules. The end result is that following modules have been designed.
Electrolyser module
Deoxer Module
Dryer Module
Water Treatment Module
Power supply and Control unit
Rack mount
The goal was to design the electrolysis system to fit into standardized racks, in order to standardize the
footprint measurements, and ease installation requirements. This goal was achieved, subsequently the only
requirement currently, is available floor space, electricity, water, and pipes for hydrogen, oxygen and
ventilation.
Demonstration (WP7)
The developed 2nd generation rack-mounted alkaline electrolyser system has been installed and
demonstrated for 18 month in a college (H2College) with 66 apartments.
A number of measurements were performed on the system installed in H2College. No measurements on
the overall system efficiency were conducted. However measurement on the stack showed a stack
efficiency of 86.5% at a current density of 177mA/cm2 and a temperature of 74.4 ˚C at the hydrogen outlet.
The system is estimated to have run roughly 2000h during the demonstration period. The demonstration of
the 2nd generation alkaline electrolyser is considered a success for several reasons. While not all technical
goals were reached, the research and development in the project have resulted in some very good
solutions for most of the challenges in high pressure alkaline electrolysis.
The concept of a rack mounted electrolyser was proven viable, and all critical aspects of the modularisation
have been covered.
The H2College system demonstrates therefor more than the development of a small size electrolyser, it is a
leap from developing electrolyser systems as a plant, to developing electrolyser systems as a product
suitable for mass production.
9
Dissemination
The project has generated an extensive amount of significant findings which have been communicated in
different ways. Besides having shared main findings from the project within reference groups under the
Danish Partnership for Hydrogen and Fuel Cells dissemination have been carried out by publications as
articles in international scientific and popular journals as well as presentations and posters on national and
international conferences. Further a patent application on the HTP cell has been filed.
For further details on the dissemination activities see chapter 11.
10
3. Background and Project Objective Energy production from wind turbines is foreseen to constitute a significant part of the future Danish
energy sources. A major challenge is however the large variation over time in wind energy production. In
order to keep production and consumption in balance load management is needed. Production of
hydrogen by water electrolysis can serve as a huge balancing load and at the same time act as an energy
storage media.
Electrolysis of water has been used in the industry for decades. Alkaline and to a certain extend PEM
electrolysis have been predominant. Characteristics for the alkaline electrolysis have been large systems
tailor-made for each plant, non-scalable and with a low efficiency. These systems are so-called 1st
generation systems.
The objective of this project has been to develop a 2nd generation alkaline system characterized by being
compact, reliable, inexpensive and energy efficient. The specific targets for the project have been to:
Increase electrode efficiency to more than 88% at a current density of 200 mA /cm2 Increase operation temperature to more than 100 degree Celsius to make the cooling energy more
valuable Obtain an operation pressure more than 30 bar hereby minimizing the need for further
compression of hydrogen for storage Improve stack architecture decreasing the price of the stack with at least 50% Develop a modular design making it easy to customize plants in the size from 20 to 200 kW Demonstrating a 20 kW 2nd generation stack in H2College at the campus of Århus University in
Herning These targets are well in line with the objectives which were established in the 2009 strategy for RD&D on electrolysis in Denmark for the period 2011-13. The objectives are summarised in the table below:
11
Table 1: The objectives for the Danish RD&D strategy for electrolysis (Strategi for Elektrolyse i Danmark, F, U & D 2010-18., 2009, Partnerskabet for Brint og Brændselsceller)
12
4. Structure of the report To pursue the targets of the project, different aspects of the electrolysis system has been in focus and
divided in seven work packages as outlined below.
4.1. Improved electrodes (WP1) There are two ways to increase the efficiency of an electrode: Increase the specific surface area per unit
and thereby get a higher current density or increase the catalytic properties of the electrode surface by
which the necessary voltage to draw the current will be lower. The work package will deal with both
through literature studies and lab tests in order to find and specify for production the optimal electrode
solution regarding efficiency and price.
To pursue the target for the project research will be carried out on three different technologies for
manufacturing and surface treatment of electrodes for improving the efficiency:
Electrochemical plating (DTU-ME)
Plasma spray with e.g. Raney Nickel at atmospheric pressure by using a protection gas (FORCE)
HTP cells made by tape casting, silk screen printing or spray painting (DTU-EC)
It is the target of this WP to reach efficiencies of at least 88% at a current density of 200 mA/cm2.
4.2. Higher operation temperature (WP2) Most commercial potassium hydroxide water electrolysers use nickel electrodes and operate at 70-80 ºC.
Only limited information on electrolysers operated at elevated temperatures (above 80°C) is available.
Increasing the operation temperature for alkaline water electrolysis from the normal 80°C will significantly
increase the efficiency. A possible obstacle for operating at elevated temperature is the lower stability of
the materials. In this project research of suitable materials for the cell design will be carried out in order to
develop the electrolysis plant for operation at elevated temperatures of more than 100 degree Celsius. This
topic is closely interconnected to the research carried out by DTU-EC on developing HTP electrodes.
Therefore, the results are reported as an integrated part of WP1 in paragraph 5.3 ‘Track 3 HTP Materials’.
(DTU-EC)
4.3. Higher operation pressure (WP3) The energy density of hydrogen is only about 1/3 of that of natural gas. Therefore the produced hydrogen
must be compressed to be used as an energy carrier. The hydrogen can be compressed in the electrolyser
or by piston compressors. The extra energy consumption used by an electrolyser operated at elevated
pressure is quite limited whereas the operation of a compressor is very energy consuming. Therefore it is
obvious to develop the 2nd generation electrolyser for operation at high pressure in order to reduce or
eliminate the demand for further compression. Elevated pressure puts special requirements on the cell
frames, gaskets and endplates (GreenHydrogen.dk).
13
4.4. Improved electrolysis stacks architecture (WP4) The electrolyser stack consist of electrodes, membranes, gaskets, ports and manifolds, tie rods and
endplates. The work package will deal with the design of the stack in order to find the optimum solution
regarding flow of electrolyte, oxygen and hydrogen, corrosion, stray currents, material use, stack volume,
suitability for mass production and price reduction. Special attention will be paid to the so-called non zero
gap concept by which it might be possible to reduce the price of the stack by a factor of 3
(GreenHydrogen.dk, CET).
4.5. Corrosion (WP5) Corrosion in alkaline electrolysers is a severe problem and increasing with increasing temperature, pressure
and stack voltage. In order to secure a lifetime of more than 10 years it is necessary to know and
understand the corrosion mechanisms. In this WP analysis and tests in connection with corrosion in hot
alkaline environment (corrosion in general as well as stress corrosion etc.) will be carried out. Explanation
of corrosion condition and the possibility of hydrogen brittleness as a consequence of hydrogen exposure
will be investigated. Material and surface characterizing by use of scanning electron microscope (SEM) and
Focused Ion Beam Microscopy (FIB). Fabrication of high pressure cells is not a part of this WP but will be
delivered by GreenHydrogen.dk (FORCE, Risø, DTU).
4.6. Modular design for easy up scaling (WP6) In order to reduce the engineering for customer design, a modular system will be developed.
Modularisation has successfully been used in the UPS industry by American Power Conversion to such
extend that the customer literately can design a new UPS system by himself. Experience from the
development of this module system will be used (GreenHydrogen.dk).
4.7. Demonstration (WP7) Demonstration will take place by substituting a 1st generation stack, which is installed at H2College, by the
2nd generation stack developed in this project. H2College consists of 66 passive houses at the campus of
Århus University in Herning powered by hydrogen made from water and wind power. The plants
performance regarding electricity consumption, hydrogen production hydrogen quality, utilisation of
cooling energy etc. will be monitored as well as operation experiences will be reported (GreenHydrogen.dk,
CET).
14
5. Improved electrodes (WP 1) Three different tracks of electrode technology have been developed in this project; Electrochemical plating
by DTU-ME, Powder Technology by FORCE Technology and HTP material by DTU-EC. The preconditions for
each technology are not alike. The electrochemical plating technology has in a PhD study at DTU-ME been
refined and developed to obtain an inexpensive method with potential for mass production in the short
term. The plasma spray technology, which is a well-known technology, has been developed with the aim to
be carried out at atmospheric pressure instead of at vacuum and has further been tested by FORCE
Technology. Finally, the third technology, where HTP materials have been used for alkaline electrolysis in
order to obtain elevated process temperature, has involved a PhD study as well. This method is still on a
research level and as such this technology track was not planned to provide results ready for use at
manufacturing level.
The differences in the three tracks regarding involvement of PhD’s as well as development level are
reflected in the reporting of WP 1 where a common standard for reporting has not been pursued.
5.1. Track 1 Electrochemical Plating
5.1.1. Literature study and preliminary experiments
Electrocatalytic activity of the hydrogen electrode
It has been shown that the best hydrogen electrode is the one having intermediate M – H bond energy (or
free energy of hydrogen adsorption ( ))[1]. When plotting the electrocatalytic activity (exchange
current density for hydrogen evolution reaction) vs. the M – H bond strength for different metals a so
called volcano plot is formed. The metals that are on the top of the volcano plot are the ones that are the
best electrocatalysts for the hydrogen evolution reaction (HER). As seen in Figure 1 platinum has the best
electrocatalytic properties for HER. Here after comes rhenium, ruthenium, iridium and silver. However, all
these metals are categorised among precious metals and therefore highly expensive. The next metals in the
plot are nickel, cobalt and iron. These three metals are therefore possible candidates as raw material for
the new hydrogen electrocatalyst to be developed.
15
Figure 1: Volcano plot of the electrocatalytic activity for hydrogen evolution reaction vs. metal – hydrogen bond of transition metals [1]
Corrosion properties
The next step for selecting proper material for the catalyst is to assure that the material is stable, does not
corrode, in the operating alkaline media. Pourbaix diagrams are commonly used in chemistry for detecting
the thermochemical stable phases of an aqueous electrochemical system. The diagrams are plotted with
the pH of the electrolyte on the x-axis and the potential of the metal on the y-axis. Figure 2 shows the
calculated Pourbaix diagrams for Iron, Nickel and Cobalt at 80°C and 1 atm. pressure. As the developed
electrodes need to be stable in strong alkaline media (pH >14) and at intermediate temperature (> 80).
Looking at the calculated diagrams at pH 14 gives an indication of the corrosion resistivity of the metals
under the expected electrolysis conditions. Between the two dotted blue lines on the diagrams water is
stable. Above the two lines oxygen is stable (oxygen evolution can take place). Whereas, below the two
blue lines hydrogen is stable (hydrogen evolution can take place). According to the diagrams, all of the
metals decompose at some point inside the -2 to 2 V potential gap. The red circles in Figure 2 indicate at
what voltage the decomposition might take place. Fortunately the thermodynamical calculations do not
affect the kinetics. That is, they do not tell anything about how fast the metals will corrode or decompose
in the particular media.
16
Figure 2: Pourbaix diagrams for Iron, Nickel and Cobalt in water at 80 °C. The red circles indicate where the metals are unstable at pH 14.
From the literature and own experience it is known that nickel decomposes particularly slowly in alkaline
media. Therefore, although nickel is unstable at some oxygen evolution potentials, this is not expected to
give problems. It is however our experience that if iron is used as an anode in alkaline media it decomposes
and deposits on the cathode. Regarding the cobalt, the literature does not entirely agree on whether it is
stable or not in alkaline media. When looking at the Pourbaix diagram for cobalt, the decomposition can
only take place inside the water stable area. Thus, according to the thermodynamics, if the electrolysis cell
is turned off electrodes made of Co might corrode. For testing this, a cobalt plate was stored in a 1 wt%
NaOH for few days thereafter the surface was inspected for corrosion. After storing a severe corrosion of
the Cobalt was observed, see Figure 3 .
Figure 3: Corrosion of Cobalt plate in 1 wt% NaOH
From the electrochemical calculations and corrosion test it was decided to select nickel as the material for
the hydrogen electrode.
Short notes about the anode
Number of papers have been published about catalysts for the oxygen evolution reaction in alkaline media
[2–8]. One of the most promising candidates not containing any precious metals are Co3O4 and Raney-
Nickel. Due to the fact that Co is not stable at zero potential and that the production cost of electrodes
would be much less if it was possible to have the same catalysts on both sides of the electrode, it was
decided to start with developing one kind of electrocatalyst for both anode and cathode. Therefore, as for
the cathode nickel was selected to be the material for the anode.
17
The effect of increasing the surface area
Electrochemical activity of a catalyst does not only depend on the intrinsic properties of the catalyst. The
structure and geometry of a catalyst also have an effect. Platinum black (or platinized platinum) is a good
example of this. Platinum black is known to be the ultimate best hydrogen catalyst having zero hydrogen
over potential in saturated hydrochloric acid whereas shiny platinum electrode is observed to have 340 mV
hydrogen over potential [9]. The reason for the large difference in electrocatalytical activity is that Pt black
has a much larger surface area compared to shiny platinum. Although, due to its price, platinum was not
considered to be a candidate for the new electrolysis electrode; but its extreme god electrocatalytic
properties was an inspiration for the development. In order to learn from the HER “master” Pt surface was
produced and its micro structure inspected. Figure 4 shows platinum black surface plated at DTU. The
platinum surface appears to be black (Black body) due to the high absorption of all incident
electromagnetic radiation, regardless of frequency or angle of incidence. Figure 5 shows images of the
platinum surface captured in a high resolution microscope, in two different magnifications. From platinum
black we have learned that the electrocatalytic activity of a surface can be largely increased by increasing
the real surface area of the catalyst and thus the amount of crystal defects, where the HER is suggested to
take place.
Figure 4: Platinum black surface electroplated at DTU
Figure 5: High resolution scanning electron microscope images of Platinum black surface plated at DTU
18
Production methods
The aim was to produce a nickel surface with large real surface area. Looking into the literature, Justi and
Winsel discovered the highly porous Raney nickel as a cathodic electrocatalyst in an alkaline media already
in 1961[10]. Raney nickel is mostly prepared by selectively leaching of aluminium from Ni-Al alloys. This
process is often referred to as activation. Lattice vacancies formed when leaching causes large surface area
and high activity due to very reactive lattice defects[11]. The activated catalysts therefore provide superior
performance, compared to smooth nickel cathodes. Plasma spraying or co-depositing nickel alloys (NiAl3,
Ni2Al3) onto a nickel or steel support are common techniques applied when producing Raney nickel
electrodes [12]. Cold rolling and hot dipping of aluminium combined with a thermo-chemical diffusion
process have also been proposed[13][14]. However, the reproducibility and durability of the Raney nickel
electrodes is often deficient[15]. The drawback of the commercial plasma spraying is that relatively brittle
nickel phase is formed. Furthermore, the adhesion between the porous surface and the underlying
substrate is often poor. From the previous it was clear that new processes for making porous nickel
structure that are robust and durable needed to be established.
Magnetron sputtering is a plasma vapour deposition (PVD) process where sputtering material is ejected
from a sputter target due to bombardment of ions to a substrate surface. In the particular case the
aluminium metal are sputtered to the electrode surface forming a thin film of aluminium. The magnetron
sputtering technique can provide superior interlayer adhesion between the substrate and the coated film.
Moreover, the thickness of the coated layer can be controlled precisely assuring unique uniformity. Thus,
instead of using the typically thermal spraying process to form a porous nickel structrue it was decided to
select the magnetron sputtering process for depositing an aluminium tinfilm onto a nickel or nickel plated
substrate and afterward form the desired Ni/Al alloys with a proper heat treatmen.
5.1.2. Electrode development - Processes for producing a stable and
durable porous nickel electrode
As for the porous nickel electrodes prepared during this project by PVD process, the effort has mainly been
concentrated on coating of commercially pure nickel. For lowering the production cost for large scale
production, plating nickel onto an iron or a stainless steel plate would be a better alternative. The process
flow for the developed porous nickel plates was as follows.
Fejl! Henvisningskilde ikke fundet., Figure 9 and Figure 10 show images of the developed nickel electrode
after Al PVD, heat treatment and leaching, respectively. After alkaline leaching the surface has a black
appearance indicating large actual surface area like a black body.
PVD of Al onto a Ni surface
Thermo-chemical diffusion process
@ 610 °C
Chemical and electrochemical alkaline leaching
of Al
19
Figure 6 Nickel plate coated with approx. 20 µm of aluminium via Magnetron sputtering is a plasma vapour deposition
Figure 7 PVD aluminium plated nickel after heat treatment
Figure 8 The PVD aluminium nickel plate after heat treatment and selectively etching of aluminium
The Al PVD was prepared on nickel surface by a non-reactive dc magnetron sputtering at The Danish
Technological Institute in Aarhus. During the thermo-chemical diffusion, the aluminium atoms diffuse into
the nickel structure and a several intermetallic phases between Al and Ni is formed. The red horizontal line
in the phase diagram in Figure 9[16] indicates which Ni-Al diffusion couples are possible at 610 °C and
atmospheric pressure. The thickness of each intermetallic phase formed depends on the amount of Ni and
Al available in the diffusion system and the heat treatment. The SEM image in Figure 9 shows cross section
of a PVD Al/Ni electrode after heat treatment prior etching.
20
Figure 9: To the left. Phase diagram for the Ni-Al diffusion couples. The red line indicates the thermo-chemical diffusion temperature selected. To the right, cross section SEM image of a PVD Al/Ni electrode after heat treatment at 610 °C. The arrows indicate the type of intermetallic phase in the PVD Al/Ni cross section.
Figure 10 and Figure 11 show cross section microscope images of the PVD Al/Ni electrodes after the heat
treatment followed by the first and second alkaline leaching procedure, PVD Al/Ni 1 and PVD Al/Ni 2,
respectively. With the first leaching process about 5 µm skeletal Ni/Al residues is formed. When selectively
leaching the Al with the second leaching procedure, the entire Al3Ni2 phase is leached, resulting in
considerably larger porous nickel layer. The difference is due to optimized leaching procedure.
Figure 10: Cross section image of a PVD Al/Ni 1. Electrode prepared by the unoptimised procedure of selectively leaching the aluminium.
Figure 11: Cross section image of a PVD Al/Ni 2 leached electrode after optimizing the selectively leaching of aluminium.
The cross section images above reveal higly porous nickel microsstructure with pores in the range of
approximately 0.5-1.5 µm. For nanostructure investigations high resolution microscope studies were
required. High resolution microscope images of the PVD Al/Ni 2 specimen surface at two different
magnifications are shown in Figure 12. When looking at the image D, where the maximum resolution is
reached, 200,000 times magnification, nanopores down to 20 nm can be detected.
21
Figure 12: HR-SEM images of the Al PVD Ni 2 surface after heat treatment and optimized alkaline etching of aluminium.
5.1.3. Electrochemical measurements
During this project cyclic sweep electrochemical methods have been used for screening of different
materials. Simultaneously, the processes capability has been examined. On the basis of the results, it was
concluded that the test setup needs to be extremely accurate. This can only be achieved with the right kind
of compensation possibilities and a high degree of care when carrying out the experiments. Due to that, a
large effort has been put into designing accurate and reliable measurement setups.
Figure 13 shows four cathodic polarisation curves recorded on a PVD Al/Ni 2 surface electrode in 1 M KOH
at 25 °C and current densities up to 1.6 V. As seen from the figure, the measurements are consistent. When
comparing the polarisation curves with measurements of smooth nickel under the same conditions, also
showed in Figure 13, it can be seen that the developed PVD Al/Ni electrodes have 400 mV lower hydrogen
over potential compared to a smooth nickel.
A B
C D
22
Figure 13: Cathodic polarization curves for PVD Al/Ni 2 repeated 4 times and smooth nickel in 1 M KOH electrolyte at 25 °C.
Figure 14 shows three anodic polarisation curves recorded on a PVD Al/Ni 2 surface. As seen from the
figure the developed PVD Al/Ni electrodes have about 50 mV lower oxygen over potential compared to
smooth nickel in 1 M KOH at 25 °C at a current up to 0.6 V.
Figure 14: Anodic polarization curves for PVD Al/Ni 2 repeated 3 times and smooth nickel in 1 M KOH electrolyte at 25 °C
23
Cathodic polarization curves made in 25% KOH at 70 °C are shown in Figure 15. The curves indicates that
the PVD Al/Ni 2 is more active towards the hydrogen evolution reaction (HER), compared to PVD Al/Ni 1.
This is in agreement with Figure 10 and Figure 11. Namely, that PVD Al/Ni 2 has larger porous nickel residue
and thus higher catalytic activity. Compared to smooth nickel, the PVD Al/Ni 1 and PVD Al/Ni 2 have around
220 and 300 mV lower hydrogen over potential at 200 mA/cm2 in 25% KOH at 70 °C, respectively.
Figure 15: Polarization curves of PVD Al/Ni 1, PVD Al/Ni 2 and blank nickel in 25% KOH electrolyte at 70 °C. Measurements made at FORCE Technology
5.1.4. Durability test
Durability test on sixteen PVD Al/Ni electrodes was carried out in a commercially produced 17-cell bipolar
electrolyser from GreenHydrogen.dk. The end plates in the electrolyser were made of pure nickel. The test
indicated no serious mitigation in the electrode efficiency during the 9000 hour testing period. The
efficiency of the electrolysis stack was measured to be around 78-79 % compared to the higher heating
value (HHV) at 65 °C. It should be mentioned that the electrolysis stack was of an old version and some
efficiency loss was caused by the stack itself and not by the electrodes. Furthermore, the first and last
electrodes in the stack were made of pure nickel. Figure 16 shows efficiency measurements made on the
stack at 22 atm. pressure and 200 mA/cm2 current density.
24
35 40 45 50 55 60 65 70 75 80
31
32
33
34
35 Stack Voltage
Efficency
Stack Temperature (0C)
Sta
ck V
olta
ge
(V
)
72
74
76
78
80
82
Effic
en
cy (
%)
Figure 16: Efficiency measurement made on a 17-cell bipolar electrolysis stack containing 16 electrodes with the developed electrocatalytic surface. The efficiency measurements were made after more than 9,000 operation hours at 22 bars and 200 mA/cm
2 current density.
5.1.5. Summary
During the presented Ph.D. study, a new type of electrodes, suitable for industrial alkaline water
electrolysers, has been developed. The electrodes are produced by magnetron sputtering a thin film of
aluminium onto a nickel or nickel plated substrate. Hereafter, the aluminium is interdiffused into the nickel
surface, by heating, and several intermetallic phases between Al and Ni are formed. Lastly, aluminium is
selectively leached from one of the intermetallic phases on the surface and high porous nickel electrode is
formed.
The structure of the developed electrodes has been characterised via high resolution scanning electron
microscopes, revealing large actual surface area with pores down to 20 nm. Consistent electrochemical
measurements show that the developed electrodes need 450 mV less voltage for decomposing water into
hydrogen and oxygen compared to smooth nickel electrodes, in 1 M KOH at 25 °C and 1 atm. durability test
in an electrolysis stack for 9,000 hours indicate no mitigation in the electrochemical efficiency of the
electrodes.
5.1.6. References
[1] S. Trasatti, “ELECTROCHEMICAL THEORY | Hydrogen Evolution,” in Encyclopedia of
Electrochemical Power Sources, Editor-in-Chief: Jürgen Garche, Ed. Amsterdam: Elsevier, 2009, pp. 41–48.
[2] C. K. Dyer, “Improved Nickel Anodes for Industrial Water Electrolyzers,” J. Electrochem. Soc.,
vol. 132, no. 1, pp. 64–67, Jan. 1985.
25
[3] H. Wendt, H. Hofmann, and V. Plzak, “Materials research and development of
electrocatalysts for alkaline water electrolysis,” Materials Chemistry and Physics, vol. 22, no. 1–2, pp. 27–
49, 1989.
[4] S. P. Singh, R. N. Singh, G. Poillearat, and P. Chartier, “Physicochemical and electrochemical
characterization of active films of LaNiO3 for use as anode in alkaline water electrolysis,” International
Journal of Hydrogen Energy, vol. 20, no. 3, pp. 203–210, Mar. 1995.
[5] H.-K. Lee, E.-E. Jung, and J.-S. Lee, “Enhancement of catalytic activity of Raney nickel by
cobalt addition,” Materials Chemistry and Physics, vol. 55, no. 2, pp. 89–93, 1998.
[6] M. El Baydi, S. K. Tiwari, R. N. Singh, J.-L. Rehspringer, P. Chartier, J. F. Koenig, and G.
Poillerat, “High Specific Surface Area Nickel Mixed Oxide Powders LaNiO3 (Perovskite) and NiCo2O4 (Spinel)
via Sol-Gel Type Routes for Oxygen Electrocatalysis in Alkaline Media,” Journal of Solid State Chemistry, vol.
116, no. 1, pp. 157–169, Apr. 1995.
[7] M. Merrill and R. Dougherty, “Metal oxide catalysts for the evolution of O-2 from H2O,”
JOURNAL OF PHYSICAL CHEMISTRY C, vol. 112, no. 10, pp. 3655–3666, Mar. 2008.
[8] J. O. Bockris and T. Otagawa, “The Electrocatalysis of Oxygen Evolution on Perovskites,” J.
Electrochem. Soc., vol. 131, no. 2, pp. 290–302, Feb. 1984.
[9] D. J. G. Ives and G. J. Janz, Reference electrodes, theory and practice. Academic Press, 1961.
[10] G. Hoogers, in Fuel Cell Technology Handbook, Taylor & Francis, 2002.
[11] V. S. Bagotsky, in Fuel Cells: Problems and Solutions, John Wiley & Sons, 2012, pp. 209–210.
[12] H. Wendt and G. Kreysa, Electrochemical Engineering: Science and Technology in Chemical
and Other Industries. Springer, 1999.
[13] “Method of producing finely-divided nickel,” U.S. Patent 162819010-May-1927.
[14] T. Boruciński, S. Rausch, and H. Wendt, “Raney nickel activated H2-cathodes Part II:
Correlation of morphology and effective catalytic activity of Raney-nickel coated cathodes,” J Appl
Electrochem, vol. 22, no. 11, pp. 1031–1038, Nov. 1992.
[15] R. H. Jones and G. J. Thomas, Eds., in Materials for the Hydrogen Economy, 1st ed., 39: CRC
Press, 2007.
[16] T. B. . B., Hugh; Bennett, L.H.; Murray, Joanne L. Massalski, in Binary Alloy Phase Diagrams.
Volumes 1 & 2. TWO VOLUME SET, 2 VOLUME SET; 1st ed., American Society for Metals, 1986, p. 142.
26
5.2. Track 2 Atmospheric Plasma Technology The activities at FORCE Technology (FT) have mainly focussed on developing new highly efficient Raney
nickel (NiAl) electrode coatings by use of atmospheric plasma spraying (APS). Such electrodes have
previously been produced by vacuum plasma spraying (VPS) /1-2/, but the APS technique represents a
much simpler and cheaper approach that possibly makes it suitable for large-scale production.
Secondly, FORCE Technology’s efforts have involved electrochemical and metallographic characterisation of
the electrode coatings produced by APS. This allowed selection and optimisation of the APS coatings before
they were produced in larger scale for the demo cell stacks at GreenHydrogen /3/.
WP 1.1-1.6 Iterative improvements of electrodes produced by APS to achieve 88 % efficiency
WP 1.7 Manufacturing of electrodes by plasma spraying
WP 1.8 Long-term tests and storage of plasma-sprayed electrodes
It was shown quite early in the project that highly efficient cathode coatings meeting the goal of 88 %
efficiency could be produced by APS. At the same time, FORCE Technology established a test concept for
characterising the electrode coatings that verified APS coatings as fully comparable or better than those
described in the literature. Both efforts placed the project right on track from the beginning.
Further developments of the APS coatings let to additional improvements in stability of substrate, adhesion
to substrate, spraying parameters and optimisation of coating thickness.
Larger electrodes in many shapes and geometries were provided for the pilot test stacks and the results
were promising. The electrodes had significantly optimized mechanical and electrochemical properties,
compared to the baseline.
Near the end of year two, only marginal improvements were made of the NiAl coatings for the cathode. At
this stage, the cathode side was considered nearly or fully developed as evidenced by the impressive
efficiency reported for the pilot cells (97 %). The subsequent development of coatings for the anode side
did not provide the desired results, but gained important insight about possible future process routes.
Today, the developed NiAl electrodes produced by APS are considered suitable for large-scale production. It
is just a matter of demand before the process can be financially attractive. This will require investments
and establishment of a company outside FORCE Technology’s auspices.
Details about the studies performed by FORCE Technology have been published in a separate report /4/.
5.2.1. Development of improved electrodes (WP 1.1-1.6)
The development of APS NiAl coatings has included iterative loops of production, testing and evaluation. In
total, approx. 35 batches of test materials have been produced.
The parameters that have been examined include:
27
Powder supplies and pre-mixing.
Addition of alloying elements (e.g. Co3O4).
Spraying parameters (on-line optimisation by plume analysis).
Coating thickness.
Substrate preparation and shapes.
Final heat treatment.
Most of the materials have been characterised by electrochemical testing and metallographic examination.
Figure 17 shows a comparison of the hydrogen evolution potential at different current densities, i.e. the
cathode performance. Compared with non-coated nickel, the NiAl coatings produced by FORCE Technology
(APS) provide a considerable improvement (i.e. reduction) in cell voltage, even at a small coating thickness
of 30 μm. The best performance is observed for NiAl 100 μm. Compared with vacuum plasma spraying
(Schiller et al.) and galvanic plating techniques, the APS technique provides better performance as cathode
material.
Figure 17. Comparison of cathode potentials at different current densities. Less numeric value means less overvoltage of the hydrogen electrode, i.e. higher efficiency
Figure 18 shows a comparison of the oxygen evolution potential at different current densities, i.e. the
anode performance. NiAl coatings produced by APS provide a considerable improvement when compared
with the reference material, non-coated nickel. However, the improvement is not as evident as for the
cathode side. Alloying with oxides like cobalt oxide (Co3O4) is necessary to obtain further improvement as
demonstrated by Schiller, who used VPS. The first attempts using APS did not provide improvement
-1350
-1300
-1250
-1200
-1150
-1100
-1050
-1000
-950
-900
-850
0 50 100 150 200 250 300 350 400 450
Hyd
rog
en
ev
olu
tio
n p
ote
nti
al (
mV
SH
E)
Current density (mA/cm²)
APS NiAl (100µm)
NiAl plated
VPS NiAl(Schiller)
APS NiAl (30µm)
Nickel (non-coated)
Perforated Ni (Schiller)
28
compared to the NiAl 100 μm coating. Consequently, further research is needed to adapt the APS
technique for producing highly efficient anode coatings.
Figure 18. Comparison of anode potentials at different current densities. Less numeric value means less overvoltage of the oxygen electrode, i.e. higher efficiency
The results presented above are based on half-cells measurements of the cathode and anode from which
the cell voltage is calculated. Guiding values of the efficiency has then been calculated by using the thermo
neutral voltage (1.48 V) as shown in Table 2. The calculated efficiency should be considered the highest
attainable in a cell without IR drop. When the coatings are installed in real electrolysers, the efficiency will
be slightly lower, depending on the IR drop in the cell’s electrolyte and membrane. However, it is clear that
a considerable improvement has been achieved with the new APS NiAl coatings, and similar results can be
expected when implementing in electrolysers as already demonstrated by GreenHydrogen.dk /3/.
Cell Voltage (V)
n=1 n=2 avg Efficiency
Nickel (non-coated) 1.975 1.940 1.958 76 %
NiAl (30µm) 1.683 1.682 1.683 88 %
NiAl (100 µm) 1.558 1.534 1.546 96 %
NiAl (300 µm) 1.544 1.548 1.546 96 %
Table 2: Guiding efficiency of NiAl coatings based on thermo neutral voltage (1.48 V). Non-coated nickel is considered as reference.
250
300
350
400
450
500
550
600
650
700
750
0 50 100 150 200 250 300 350 400 450
Ox
yg
en
ev
olu
tio
n p
ote
nti
al
(mV
SH
E)
Current density (mA/cm2)
Nickel (non-coated)
NiAl plated
APS NiAl (30µm)
Perforated Ni (Schiller)
APS NiAl (100µm)
Co3O4/NiAl mix (Schiller)
29
5.2.2. Manufacturing of electrodes by plasma spraying (WP 1.7)
The systematic screening of Atmospheric Plasma Spraying parameters using a plume analyser combined
with analysis of bond strength, pre-treatment and stand-off resulted in un-deformed 0.5 mm nickel plates
with a surface area equivalent to 600 cm2. The bond strength of the coating was defined as minimum 57
MPa, when not activated. Furthermore, the NiAl consumption was reduced by 74 %.
A major problem was the tension induced by blasting, which was released under the coating process and
caused extensive deformation. A method of applying sufficient surface roughness to substrate with minimal
adding of tension and deformation was examined. The study suggested that the best results were obtained
by combining annealing and sandblasting, when the objectives were minimum deformation and maximum
adhesion. However sanding could be sufficient when mesh metal, as the substrate, was to be coated. The
applied tensions in the mesh metal are more easily released without causing deformation.
The abovementioned practical problems related to manufacturing NiAl coatings by APS have been
identified and solved during the project. The gained experience may be transferred to other cost-effective
systems that are cheaper than the nickel substrates used in the project till now.
5.2.3. Long-term tests and storage of plasma-sprayed electrodes (WP 1.8)
This work package was scaled down due to higher priority of other activities, i.e. testing of polymer
materials. A few tests have been carried out, mainly to evaluate requirements for storage/handling and
long-term behaviour of NiAl electrodes.
In most cases, the cathode efficiency decreased while the anode efficiency was unaffected after storing
NiAl coatings for long time. Air and water seem as a better medium for storage than KOH. NiAl coatings are
fragile and conditioning is decisive for the performance. This is believed to have caused larger spread in the
results. The study indicated that procedures for handling and conditioning of NiAl coatings are more crucial
than the actual storage media.
The obtained results of extended polarisation of NiAl coatings at -200 mA/cm2 were not conclusive. In one
case, a large decrease is observed in cathode performance over time while another specimen showed good
performance. Further refinement of the lab test method is required to obtain consistent results. Promising
results on long-term performance of selected APS NiAl coatings have been obtained in pilot cell stacks at
GreenHydrogen /3/.
5.2.4. References
1. Schiller, G. Henne, R. Mohr, P. and Peinecke, V., High Performance Electrodes for an Advanced Intermittently Operated 10-kW Alkaline Electrolyser, Int. J. Hydrogen Energy, Vol. 23, No. 9, pp. 761-765, 1998. 2. Schiller, G. Henne, R. and Borck, V., Vacuum Plasma spraying of High-Performance Electrodes for Alkaline Water electrolysis, JTTEE5, Vol. 4, No. 2, pp. 185-194, 1995. 3. Final Report, Second Generation Alkaline Electrolysis, EUDP08-II, GreenHydrogen. 4. Troels Mathiesen, Peter Tommy Nielsen and Lisbeth R. Hilbert; Second Generation Alkaline Electrolysis, Final Report of FORCE Technology, 109-22486/TRM & 109-22502/PTN, 29 June 2012.
30
5.3. Track 3 HTP Materials
5.3.1. Abstract of Frank Allebrod Ph.D. thesis “High temperature and
pressure alkaline electrolysis”
Alkaline electrolysers have proven to operate reliable for decades on a large scale (up to 160 MW), but in
order to become commercially attractive and compete against conventional technologies for hydrogen
production, such as natural gas reforming, the production and investment costs have to be reduced. A
reduction of the investment costs may be achieved by increasing the operational pressure and temperature
of the electrolyser, as this will result in: 1) production of pressurized hydrogen and oxygen, 2) improved
electrical efficiencies and 3) increased current density, i.e. increased hydrogen production rate for a given
electrolyser cell area.
This thesis describes an exploratory technical study mainly in order to examine the possibility to produce
hydrogen and oxygen with a new type of alkaline electrolysis cell at high temperatures and pressures. To
perform measurements under high pressure and at elevated temperatures it was necessary to build a
measurement system around an autoclave, which could stand high temperatures up to 250 °C and
pressures up to 95 bar as well as extremely caustic environments. Based on a literature study to identify
resistant materials for these conditions, Inconel 600 was selected among the metals which are available for
autoclave construction. An initial single atmosphere high temperature and pressure measurement setup
was build comprising this autoclave. A second high temperature and pressure measurement setup was
build based on experiences from the first setup in order to perform automatized measurements.
The conductivity of aqueous KOH at elevated temperatures and high concentrations was investigated using
the van der Pauw method in combination with electrochemical impedance spectroscopy (EIS). Conductivity
values as high as 2.7 S cm-1 for 35 wt%, 2.9 S cm-1 for 45 wt%, and 2.8 S cm-1 for 55 wt% concentrated
aqueous solutions were measured at 200 °C. The conductivity of immobilized KOH was determined by the
same method in the same temperature and concentration range. Conductivity values as high as 0.67 S cm-1
for 35 wt%, 0.84 S cm-1 for 45 wt%, and 0.73 S cm-1 for 55 wt% concentrated immobilized aqueous solutions
were determined at 200 °C.
A new type of an alkaline electrolysis cell was developed in order to operate at high temperatures and
pressures. Aqueous potassium hydroxide immobilized electrolyte in porous SrTiO3 was used in those cells.
Electrolysis cells with metal foam based gas diffusion electrodes and immobilized electrolyte were
successfully demonstrated at temperatures up to 250 C and 40 bar. Different electro-catalysts were
tested in order to reduce the oxygen and hydrogen overpotentials. Current densities of 1.1 A cm-2 and 2.3
A cm-2 have been measured at a cell voltage of 1.5 V and 1.75 V, respectively, without using expensive
noble metal catalysts. Electrical efficiencies of almost 99 % at 1.1 A cm-2 and 85 % at 2.3 A cm-2 combined
with relatively small production costs may lead to both reduced investment and operating costs for
hydrogen and oxygen production. One of the produced electrolysis cells was operated for 350 h. Based on
the successful results a patent application covering this novel cell was filed. Assuming that the developed
cells will be scaled up and successfully tested for some thousand hours, they may offer an important role in
future energy storage scenarios.
31
5.3.2. Important literature data
Thermodynamics of water splitting
There are significant advantages of performing water or steam electrolysis at elevated temperature. The
following section describes how the reversible cell voltage and the thermo-neutral voltage behave as a
function of temperature and pressure. Furthermore, methods to calculate these potentials are given that
make it possible to calculate those without the necessity to use steam tables.
The thermo-neutral voltage Etn describes the potential, at which the process in the electrolysis cell is
neither endothermic nor exothermic; the electrical losses equal the heat required by the endothermic
electrolysis process, and describes the total energy demand for the reaction. It can be calculated by
equation (1), with ΔHf being the enthalpy of formation of water or steam, n is the number of electrons
involved in the reaction, and F is Faraday’s constant (96485 C∙mol − 1).
1f
tn
HE
nF
The thermo-neutral voltage Etn,0 at standard temperature and pressure (STP, 25 °C, 101.325 kPa) is 1.481 V
for ΔH0f = − 285.840 kJ mol − 1. The enthalpy for water splitting can be calculated using enthalpies from
steam tables for temperatures above 25 °C.
Etn as a function of temperature at standard pressure is shown in Figure 19. It must be pointed out that
parts of the figure are only theoretical, since water at standard pressure and temperatures above 100°C
becomes steam and steam will partly condensate to water below 100 °C, respectively. The theoretical
values are therefore given as dashed lines. The influence of the pressure to the thermo-neutral voltage is
negligible for most practical applications. The numerical change is only 2.2 mV for a pressure change of 100
bar. It will therefore not be calculated explicitly.
Figure 19. Thermo-neutral voltage Etn as a function of temperature at standard pressure calculated by empirical equations given by LeRoy et al.
32
The reversible (or equilibrium) cell voltage, Erev, is the minimal voltage at which electrolysis of H2O is
possible and can be calculated by equation (2), where ΔGf is the Gibbs free energy of the reaction, n is the
number of electrons involved in the reaction, and F is Faraday’s constant. At standard conditions it is 1.229
V with ΔG0f = −237.178 kJ mol − 1.
(2)f
rev
GE
nF
A comparison of the electrical energy demand, the total energy demand and the heat demand for water
electrolysis at standard pressure is shown in Figure 20 with thermodynamic data obtained from FactSage, a
commercial electronic data package and calculation program. It is shown that the electrical energy demand
for water electrolysis decreases with increasing temperature, while heat demand rises and the total energy
demand is almost constant both for the liquid and the gaseous phase.
As it is not common to perform electrolysis at standard conditions, it is necessary to know both the
influence of pressure and temperature to the reversible cell voltage. The reversible cell voltage as a
function of temperature and pressure, Erev(T, p), can be calculated by equation (3), where R is the Gas
constant (8.3144621 J mol − 1 K − 1), aH2 is the hydrogen activity, aO2 is the oxygen activity, aH2O is the activity
of water and T is the temperature in Kelvin.
2 2
2
10 2
( , ) ln· ·
3H Of f
rev T p
H O
a aG G RTE
n F n F nF a
33
Figure 20 Temperature dependence of the total energy demand, ΔHf (full line), the electrical energy demand, ΔGf (dashed line), and the heat demand T⋅ΔSf (dotted) for water (T ≤ 100 °C) and steam (T ≥ 100 °C) electrolysis at 1 bar using data from FactSage.
Erev(T, p) can also be obtained from steam tables or CALPHAD (Computer Coupling of Phase Diagrams and
Thermochemistry) programs like FactSage, which include data for ΔGf at elevated temperatures and
pressures. Figure 20 reveals that if surplus heat is available to raise steam at the pressure of the
electrolyser then a gain in electrical efficiency of several percent is obtained. At 1 bar it amounts to 13 %
electrical energy that may be substituted with thermal energy. Furthermore, Figure 20 reveals that an
electrolyser operated near above Etn is self-cooling, i.e. no cooling arrangement is necessary. The cell
voltage has to be slightly above Etn in order to compensate for heat loss to the surroundings. The size of
heat loss is naturally dependent on the degree of insulation of the electrolyser.
Solubility of oxygen and hydrogen in aqueous KOH and water
Tromans developed a formula to predict the oxygen solubility in different inorganic solutions. This formula
is said to be feasible to predict the oxygen solubility in potassium hydroxide solutions in a wide range of
pressures, temperatures and concentrations. The comparison of his formula with available data showed
satisfying accuracy up to 100 °C, but for higher temperatures no data are available.
Figure 21 shows the variation of the oxygen solubility in potassium hydroxide solution as a function of the
temperature and different concentrations of KOH (aq). It is seen that the solubility is decreasing strongly
with increasing concentration. It is also seen that the solubility has a minimum at around 100 °C. The
prediction was shown to be valid for temperatures up to the boiling point of the solution at ambient
pressure. Comparison of Tromans formula with data for other solutions that he described in his work, i.e.
NaOH, indicates that his formula is accurate and can be used also for temperatures above 100 °C.
34
Figure 21 Oxygen solubility as a function of temperature in water (o) and aqueous solutions of KOH for a concentration of 15 wt% (□), 25 wt% (x), 35 wt% (+), and 45 wt% (∇). The solubilities were calculated after an empirical formula developed by Tromans.
Hydrogen seems according to the literature to be in the order of 30 % less soluble than oxygen in aqueous
KOH at 30 C. The pressure dependence of the O2 and H2 solubility follows the simple equation (4), where
ci(T, Pi) is the solubility of the species i as a function of temperature and pressure, ci(T) is the solubility of the
species i as a function of temperature, Pi is the partial pressure of the species i above the solution and Pi, 0
is the pressure at which the data were obtained.
c i(T , Pi) = c i(T) P i P − 1
i , 0 (4)
Experiments have shown that the solubility of oxygen and hydrogen in aq. KOH as a function of the
temperature have minimum around 100 °C, where the solubility is approx. two times lower than at 25 °C.
At 250 °C it was shown that the solubility increases to values more or less two times higher than at 25 °C for
the O2 solubility in water and KOH solutions. The H2 solubility in water is approx. four times higher at 250
°C than at 25 °C. The solubility of oxygen and hydrogen as a function of the KOH concentration decreases
strongly with increasing concentration. The O2 and H2 solubility in 35 wt% KOH is approx. 21 and 30 times
lower, respectively, than that of pure water. Since no experimental solubility data for O2 and H2 in KOH
(aq) are available above 100 °C, the theoretical values must be taken with care until they are validated
experimentally.
35
5.3.3. Constructed measurement set-ups
Three different measurements set-ups were constructed as part of this project. 1) An atmospheric pressure
three-electrode set-up in order to get quickly started and to compare results with other project partners. 2)
A single atmosphere pressurized set-up and 3) a fully automated set-up with a cell house for test of 4 cells
at a time with different (the relevant) atmospheres on each electrode.
Atmospheric pressure three-electrode set-up
The 1 bar pressure set-up was built in order to study half-cell reactions of electrochemical cells. For this
purpose it is common to use a so called 3 electrode setup, which consists generally of a working electrode
(WE), a reference electrode (RE) and a counter electrode (CE). The WE is the electrode of interest. It is
important to perform measurements under a defined atmosphere; a sufficient amount of gas stream
through the electrochemical cell is therefore necessary and can for example be realized as shown
schematically in Figure 22: Argon or air can be flushed through a beaker in order to wash and humidify the
gas before it enters the electrochemical cell. Furthermore, continuous gas flow ensures that no explosive
gas mixtures will occur in the cell beaker from hydrogen and oxygen evolution. The electrical contacts of
the WE, RE and CE are connected to a potentiostat in order to perform electrochemical measurements like
cyclic voltammetry (CV) or electrochemical impedance spectroscopy (EIS).
Figure 22 Schematics of the assembly of an atmospheric 3 electrode setup for electrochemical measurements in aqueous solutions. The working electrode (WE), the reference electrode (RE) and the counter electrode (CE) are connected to a potentiostat in order to perform electrochemical measurements. Argon or air can be flushed through a beaker (right beaker) in order to wash and humidify the gas before it enters the electrochemical cell (left beaker). A heated oil bath allows measurements at elevated temperatures
36
Single atmosphere high temperature and pressure set-up
Next a single atmosphere high temperature and pressure set-up was built. For this purpose it was
necessary to build a measurement system around an autoclave which could stand high temperatures up to
250 °C and pressures up to 95 bar as well as extremely caustic environments. PTFE and PFA would be
feasible for the high temperatures, but not for the high pressures. A metallic autoclave had to be build or
ordered after a careful selection of construction materials. Inconel alloy 600 is recommended for alkalis
from Special Metals Corporation, but it is recommended to anneal the alloy if stress corrosion cracking is an
issue. Anyway, Inconel 600, also called Alloy 600, showed the most promising resistances among the metals
which are available for autoclave construction. After comparing the variety of choices on the market and
obtaining quotations from different suppliers, a Parr autoclave Type 4760 was ordered. The technical
specifications are listed in Table 3.
Producer Type Vessel
material
Volume Pmax Tmax Liner Tmax with
liner
Parr
Instruments
4760 Inconel
600
600 ml 200 bar 350 °C PTFE 250 °C
Table 3 Technical specifications of the Parr autoclave
To perform electrochemical measurements, the autoclave had to be equipped with additional gas handling
components and fittings as well as wire feed-throughs. The complete setup is shown in Figure 23; A gas
bottle (here shown as nitrogen) is used for the pressurization of the autoclave. A pressure reduction valve
PrR1 was installed to avoid too high pressure increase rates and over-pressurization of the system during
initial pressurization. The back-flow protection valve BFP1 avoids the back-flow of hazardous gases in the
gas bottle. The pressure is adjusted by manually stopping the flow of pressurized gas to the autoclave by
needle valve NV1. The J-Type thermo-well T1 measures the Temperature in the autoclave. The pressure
gauge PG1 displays the actual pressure analogue in order to allow steady control, even if the electricity
supply to the system fails, while another pressure gauge allows recording the pressure together with the
temperature in the temperature controller unit. A Teflon liner TL1 protects the inner part of the autoclave
from corrosion.
37
Figure 23 Schematic description of the measurement setup: PrR1= Pressure reducer, BFP1= Backflow protection, NV1= Needle valve, T1= J- Type thermo-well, PG2= Pressure gauge, TL1= Teflon liner, SCMC= Symmetrical Cell Sample Holder, ICMC= Immobilized conductivity measurement cell, CB1& CB2= Connection boxes, PS1 & PS2= Gamry Type 600 potentiostat.
The sample holders, for example the aqueous conductivity measurement cell ACMC, was connected to the
protection hood of the thermo-well by a PTFE shrinking tube in the center of the autoclave. The
immobilized conductivity measurement cell ICMC was situated directly below and equally fixed. Other
sample holders were fixed in a comparable way. The connection wires from each cell were put through to
the outside of the autoclave by high pressure glands (the wire feed-throughs) and were then connected to
the connection boxes CB1 or CB2. A Gamry Type 600 or 3000 potentiostat, PS1 and PS2, was used to
perform the electrochemical measurements. An aqueous solution of potassium hydroxide with lower
concentration of than the sample being measured was placed at the bottom of the Teflon liner to avoid
drying of the samples during the measurements. Figure 24 shows a sketch of a convenient sample holder
used in this setup.
38
Figure 24 Principal cross section of the Alkaline Electrolysis Cell Holder (AECH): The electrolysis cell is centered in the sample holder. A piece of metal foam acts as current collector. The Inconel tubes press the foam against the active electrode of the electrolysis cell. Nitinol springs are assembled to provide constant force in order to avoid contacting problems. They are insulated against the screws by a PTFE O-ring. The wires WE, WS, CE and Ref indicate how the cell was connected to the potentiostat. The Pt-wire acts as reference electrode in the three electrode setup.
Automated high temperature and pressure setup
The measurements which were performed in the Parr autoclave as described above were time limited to
avoid the formation of explosive gas mixtures in the reaction chamber. In order to perform measurements
over an extended period of time it was necessary to construct a system which is able to supply the reaction
chamber with a continuous flow of H2, O2, N2 and steam. Since it was desired to have an automated
operation of the system at temperatures of up to 300 °C and pressures of up to 95 bar, a number of safety
issues had to be considered. The design and construction work were done by specialized employees in
order to fulfill the safety demands. The completed two systems with the internal identification names “Rig
28” and “Rig 29” (a similar system in which the Parr autoclave was implemented eventually) are shown in
Figure 25 along with the external gas supply panel and the mobile unit for data acquisition.
39
Figure 25 Test rigs 28 and 29 for automated electrochemical tests in the temperature range from room temperature to 250 °C and pressures up to 95 bar. The mobile unit for data acquisition consists of two Gamry type 3000 potentiostats, a 8 channel multiplexer and a personal laptop computer.
A detailed diagram comprising the used actuators is shown in Figure 26. The gases enter the autoclave
through the inlets 3, 5 and 7 in Figure 26. N2, CO2 and O2 are fed through one tube (inlet 3) and H2 is fed
through inlet 7 to the catalytic burner (CB). The CB comprises a honeycomb structured ceramic body with a
catalyst to allow steam generation by H2 and O2 combustion in the autoclave with a known flow rate,
controlled by mass flow controllers. An excess of H2 or O2 can be programmed for the case that
measurements should be performed in reducing or oxidizing atmosphere. A second catalytic burner placed
right below the cell holder and before the outlet of the autoclave, ensures that the H2 and O2 produced by
electrolysis will recombine into H2O before exiting the autoclave. (The shown mass spectrometer (MS) and
the heated CO2 and transfer lines are used for other projects that have contributed to the construction of
these rigs).
40
Figure 26 Rig 28 gas handling system
Figure 27 Illustration of the sample holder used in the autoclave to test the electrolysis cells with mounted sample and simplified demonstration of the gas flow. FobAEC is the abbreviation used for (metal) foam based alkaline electrolysis cell.
41
The sample holder used to test the produced electrolysis cells is schematically shown in Figure 27. A K-type
thermocouple is used to connect the cathode of the metal foam based alkaline electrolysis cell, FobAEC, to
the counter electrode (CE) and reference input (Ref) of the potentiostat multiplexer. A tubular, headless
screw is used to fix the setup initially. A nickel spring ensures that the nickel mesh, which is wrapped
around a Teflon tube, is pressed to the cathode during the measurements. A Teflon ring has been placed
around the FobAEC to avoid gas mixing. A layer of nickel foam is used as current collector at the anodic side
of the FobAEC. Another K-type thermocouple is connected to the nickel foam current collector on the
sample bottom. It is connected to the working electrode (WE) and the working sense (WS) of the
potentiostat. The same connections were used on the right sample shown in Figure 27, but have not been
drawn for clarity reasons.
The gas mixture from the catalytic burner CB at the gas inlet (in the experiments discussed later on a
mixture of H2, H2O and N2) flows partly to the cathode of the FobAEC and further through the autoclave.
The oxygen produced at the anode of the FobAEC escapes the sample holder through a tubular hole, where
it will mix with the other gases and react to form steam at the CB placed at the autoclave outlet. A pressure
control valve is placed at the outlet stream of the autoclave to set the gas pressure to the desired value.
5.3.4. Specific electrical conductivity measurements on free and
immobilized aqueous KOH
Figure 28 Measured conductivity and cubic regression analysis for 35 wt% (o = measured, dash-dotted line = regression), 45 wt% (□ = measured, full line = regression) and 55 wt% (+ = measured, dashed line = regression) immobilized KOH (aq) as a function of temperature and a pressure of 25-40 bar.
42
The conductivity of aqueous KOH at elevated temperatures and high concentrations was investigated using
the van der Pauw method in combination with electrochemical impedance spectroscopy (EIS). Conductivity
values of 2.7 S cm-1 for 35 wt%, 2.9 S cm-1 for 45 wt%, and 2.8 S cm-1 for 55 wt% concentrated aqueous
solutions were measured at 200 °C. Micro- and nano-porous solid SrTiO3 pellets were produced and used to
immobilize aqueous KOH solutions. These are intended to operate as the ion-conductive electrolytes in
alkaline electrolysis cells, offering high conductivity and corrosion resistance. The conductivity of
immobilized KOH has been determined by the same method in the same temperature and concentration
range. Conductivity values as high as 0.67 S cm-1 for 35 wt%, 0.84 S cm-1 for 45 wt%, and 0.73 S cm-1 for 55
wt% concentrated immobilized aqueous solutions were determined at 200 °C. Furthermore, phase
transition lines between the aqueous and aqueous + gaseous phase fields of the KOH/H2O system were
calculated as a function of temperature, concentration and pressure in the temperature range of 100-350
°C, for concentrations of 0-60 wt% and at pressures between 1 and 100 bar. Figure 28 gives the data for the
immobilized electrolytes.
5.3.5. Alkaline electrolysis cell for 250 C and 40 bar
A new type alkaline electrolysis cell and method to produce them has been developed based on inspiration
from literature and a long “trial and error” process. We call it metal foam based alkaline electrolysis cell,
FobAEC, even though several variants of the cell type have been produced and tested as described in this
section.
Principle of metal foam based alkaline electrolysis cell with immobilized KOH
A porous structure in which the liquid electrolyte is immobilized by capillary forces can be used as a
combination of diaphragm and electrolyte. In this case, the reduced free volume for the liquid electrolyte
and the tortuosity of the porous structure are expected to result in a decrease in conductivity, and by that
to an increase in ohmic losses. Nevertheless, the ohmic losses can be reduced by reducing the thickness of
the porous structure. A great advantage of immobilizing the electrolyte is the possibility to use gas diffusion
electrodes (GDEs) for steam electrolysis at temperatures well above the boiling point of water, i.e. up to
250 °C in case of pressurization, while the electrolyte remains in the liquid phase. At any given
temperature, it is important to operate within an appropriate pressure range since too high pressures could
cause condensation of steam (and thereby dilute the electrolyte and/or flood the GDE) while too low
pressures could cause evaporation of the electrolyte (through which ohmic losses would increase).
Gas diffusion electrodes allow in principle electrolysis without the formation of gas bubbles, which usually
leads to increased ohmic losses due to area reduction and increased anodic and cathodic overpotentials
due to reduced available electrode area. GDEs have to be highly electronic conductive to minimize ohmic
losses, and porous to allow the produced gases to escape while steam flows into the cell towards the triple
phase boundary, TPB, as shown in Figure 29, which illustrates the FobAEC. Since the electrochemical
reactions take place at or near the TPB only, it is important that the TPB-length is as long as possible. Nickel
is known to be corrosion resistant in alkaline media and electro-catalytically active, hence it is often used as
a base material for electrodes in alkaline electrolysis cells. Electro-catalysts are optimally placed at the TPB
in order to achieve low overpotentials. Silver is known to be an excellent catalyst for the oxygen reduction
43
reaction, ORR, in alkaline fuel cells and can be considered as catalyst for the oxygen evolution reaction,
OER, as well. Co3O4 and Co- based spinel oxides are also known to be good electro-catalysts for the OER,
and Raney-nickel and molybdenum are preferred electro-catalysts for the hydrogen evolution reaction,
HER. The anodic overpotential caused by the OER is usually dominating over the cathodic overpotential
caused by the HER.
Nickel foams are widely used and have already been shown to work as a gas diffusion electrode for the ORR
in an alkaline fuel cell. More advanced metal foams have recently been developed from Alantum Europe
GmbH, where metal alloy foams are obtained by a powder metallurgical process. Important material
parameters, i.e. final composition, the specific surface area, and the pore size of the foam can be
customized to the application.
Figure 29 Illustration of an alkaline electrolysis cell with KOH (aq) electrolyte immobilized in a porous matrix and gas diffusion electrodes.
The foam based alkaline electrochemical cells (FobAEC) were produced from metal foams delivered by
Alantum Europe GmbH. Circular foam pieces with a diameter of 12 mm were stamped out of the foam
sheets with a thickness of 1.0 mm for nickel foam, and 1.6 mm for Inconel 625 foam (Inconel 625 alloy
composition is for example in wt%: Cr, 21.32; Mo, 8.58; Nb, 3.73; Al, 0.18; Ti, 0.16; Fe, 0.11; Si, 0.09; C,
0.053; Mn, 0.04; Mg, 0.01; Balance Ni). The circular foam pieces were cleaned in ethanol; beside that, the
foams were used as delivered. One or two foam layers are placed in a uni-axial press and 0.25-1.5 g SrTiO3
powder is placed on top of the foam. One drop of an appropriate binder, i.e. MEK in 33 % ethanol, has been
mixed to 100 mg of the perovskite SrTiO3 powder to obtain better adhesion of the powder particles during
processing and sintering. Another one or two layers of foam is placed on top of the powder. The whole
structure is pressed for 30 s with a pressure of 13.0 kN cm-2. The cells have then been sintered in air at 450
°C for 2 h with a heating ramp of 100 K h − 1 followed by a second sintering step in 9 % H2 in Ar at 1000 °C for
6 h with a heating and cooling ramp of 50 K h − 1.
44
Electrodeposition of silver was performed onto some of the cells in order to improve the performance of
the anode for the OER. The electrodeposition was performed with a 0.1 M AgNO3 + 1 M KNO3 solution
under N2 atmosphere and stirring. Pulsed deposition with pulse/pause relation of 1s ⁄ 1s and a potential
difference of -3.0 V (vs. a platinised Pt counter electrode) has been used to deposit the Ag particles on the
foam with a Gamry type 600 or 3000 potentiostat. Scanning electron microscopy SEM was used to analyse
the surface and cross section of the FobAEC before and after measurements. SEM (Scanning electron
microscopy) images of the untreated foam are shown in Figure 30. The pore size of the nickel and Inconel
foam is ca. 450 μm with a porosity of 95 % for the nickel foam and > 91 % for the Inconel foam according to
the manufacturers’ specification. It can be seen that the surface of the nickel foam is relatively smooth
compared to this of the Inconel foam. The coarse layer on the Inconel foam shows the Inconel layer, which
has been coated on nickel foam (left part of Figure 30).
Figure 30 SEM picture of untreated nickel foam (upper-left and lower-left) and Inconel foam (upper-right and lower-right). Note: the magnifications are different as indicated by the scale bars.
45
Figure 31 SEM picture of the surface of the nickel foam anode of a FobAEC at low (left), and high magnification (see scale bar) showing the interface between the nickel foam and the SrTiO3 (right).
The porosity of the foam is significantly reduced after pressing of the whole cell body in the uni-axial press.
The typical pore size diameter is reduced to values in the order of 150 μm ± 50 μm with an even
distribution. The shape is naturally changed to a more flat form instead of being ball-like, as it can be seen
in Figure 31, left image. The highly porous SrTiO3 structure sticks out of the nickel foam at several places up
to the surface of the cell. It can be seen that the nickel foam can act as a current collector as well as the
active electrode in such structure. The right image of Figure 31 shows the interface of the SrTiO3 and the
nickel foam, which will act as the TPB after immersion of the cell into liquid electrolyte. The SrTiO3 has a
tendency to form clusters on the surface which have a size of ca. 2 − 10 μm, while the pore size is mainly
distributed around 60 nm.
Silver electro-deposition was successfully applied to the nickel foam anode as shown in Figure 32. The left
image shows the foam with silver crystals of a particle size in the region of 10 − 20 μm and nano sized silver
particles distributed all over the foams surface. The right image shows the interface of the same sample at
higher magnification where a relatively large silver particle of 7.5 μm can be seen in the upper left part with
well distributed nano sized particles in the range of 50 − 500 nm.
46
Figure 32 SEM picture of the surface of the nickel foam after Ag electrodeposition of the anode of a FobAEC at low (left) and high magnification (see scale bar) (right) showing the interface between the nickel foam and SrTiO3 as well as the nano sized Ag particles
The results of the cross section analysis are shown in Figure 33. The brightest structure indicates the metal
foams, light grey is the porous SrTiO3 structure and dark grey is the surrounding epoxy coating for the SEM
analysis. Both electrodes are pressed out of two layers of the corresponding metal foam. The Inconel foam
is less compressed than the nickel foam and shows a higher specific surface area than the pure nickel. It can
also be seen that it will be sufficient to use only one layer of Inconel foam, but 2 layers of nickel foam are
needed in order to obtain a good distribution of gas channels in combination with the ability of the foam to
work as a current collector and active electrode.
Figure 33 SEM picture of the cross section of a FobAEC with two layers of nickel and two layers of Inconel foam showing the nickel anode (left), the Inconel cathode (center) and the full cell (right).
Electrochemical performance
The electrochemical testing comprised the 1 bar test of plain Ni and Raney-nickel electrodes for
comparison of measurement techniques between the project partners and for creating a kind of reference
against which the high pressure and temperature tests could be evaluated. Testing of the FobAEC cell type
with different electrocatalysts was carried out in the next step.
47
Raney-nickel electrodes at different temperatures
A Raney-nickel sample delivered by FORCE-Technology was tested in the above described atmospheric
measurement setup. After activation in 30 wt% KOH with 10% KNa-tartrate tetrahydrate at 80 °C for 24 hrs
(Step1), the sample has been conditioned at − 10 mA cm − 2 for 2 hours (Step 2) and at − 200 mA cm− 2 for
14 hours (Step 3) in 25 wt% KOH (aq.) after instructions from FORCE Technology.
Figure 34 Cyclic Voltammograms of the Force Raney Nickel electrode under N2 atmosphere at ambient pressure and temperatures from 70 °C to 118 °C in 10 M KOH (aq) measured with a sweep rate of 10 mV s
− 1.
Cyclic voltammetry was applied to the activated and conditioned electrode at temperatures of 70 °C, 100 °C
and 118 °C. The obtained results are shown in Figure 34. The potential for the OER at a current density of
200 mA cm− 2 was found to be 0.77 V, 0.6 V and 0.51 V for temperatures of 70 °C, 100 °C and 118 °C,
respectively. Hence, the anodic polarization was reduced by 260 mV through heating from 70 °C to 118 °C.
The potential for the HER at a current density of − 200 mA cm − 2 was found to be -1.52 V, -1.54 V and -1.58
V for temperatures of 70 °C, 100 °C and 118 °C, respectively. All measurements were carried out without IR-
compensation.
Test of FobAEC at high pressure and temperature
Cell performance tests have mainly been performed at temperatures in the region of 250 °C and pressures
around 40 bar. Cyclic voltammograms of different compositions of the FobAECs are shown in Figure 35. The
current density at a polarization of Ucell = 1.5 V (uncompensated for ohmic resistance) reached from 466 mA
cm − 2 at 240 °C and 38 bar (for a cell with 2 layers of nickel foam as anode and two layers of Inconel foam as
cathode), to 1000 mA cm − 2 at 240 °C and 37 bar (for a cell with 1 layer of nickel foam with Ag- deposition
as anode and one layer of Inconel 625 foam as cathode).
48
Figure 35 IV- curves of different cells at temperatures around 240 °C and pressures around 40 bar, i.e.: Cell 3a (x) and Cell 3b (full line) with 1 layer Ag- deposited nickel anode and 1 layer Inconel as cathode; Cell 2 (+) with 1 layer nickel foam anode and 1 layer Inconel as cathode; Cell 1b (□) with 2 layers nickel foam anode and 2 layers Inconel as cathode; Cell 1a (o) with 2 layers nickel foam anode and 2 layers Inconel as cathode.
Additional performance data for further cell compositions are shown in Table 4. The serial resistance Rs has
been measured with EIS under polarization of 1.4 V. The Ag- activated cells have shown the best
performance as shown in Table 4.
Impedance analysis have generally been performed at a polarization voltage of 1.4 V. Figure 36 shows the
cyclic voltammogram for a Ag- activated cell along with polarized impedance measurements before and
after recording the voltammogram at polarization voltages of 1.4 V and 1.5 V. The serial resistance Rs
decreased slightly after recording the voltammogram from 133 mΩ cm2 to 125 mΩ cm2 at 1.4 V
polarization, while the area specific resistance of the cell, ASRcell, decreased from 335 mΩ cm2 before, to
319 mΩ cm2 after recording the voltammogram. The Rs at 1.5 V was equal to this at 1.4 V (125 mΩ cm2),
whereas the ASRcell reduced to 250 mΩ cm2. The main change is associated with the low frequency part of
the spectrum indicating an increased gas transport resistance with time.
Cell Anode Cathode
Temperature
[°C]/Pressure
[bar]
I [mA cm −
2] at 1.5 V
(uncomp.)
I [mA cm −
2] at 1.75 V
(uncomp.)
Rs [mΩ
cm2] at 1.4
V
polarization
U
cell
[V] at 0.5 A
cm − 2
(comp.)
Cell 1a 2 layer Ni
foam
2 layer Inc
foam 240/38 466 1130 148 1.441
49
Cell 1b 2 layer Ni
foam
2 layer Inc
foam 237/36 528 1378 140 1.419
Cell 2 1 layer Ni
foam
1 layer Inc
foam 240/39 679 1570 112 1.388
Cell 3a
1 layer Ni
foam + Ag
deposition
1 layer Inc
foam 240/37 1000 1918 119 1.291
Cell 3b
1 layer Ni
foam + Ag
deposition
1 layer Inc
foam 247/42 903 1978 124 1.337
Table 4 Comparison of the cell performance for different compositions of the FobAEC at around 240 °C and 40 bar.
Figure 36 Cyclic voltammogram (upper figure) of a FobAEC with Ag- deposited anode and Inconel cathode at 247 °C and 43 bar in comparison with the corresponding impedance analysis (central figure) at polarization voltages of 1.4 V before (x) and 1.4 V (o) and 1.5 V (∇) after the cyclic voltammetry test.
50
The performance of a FobAEC with Ag- deposited anode and Inconel cathode at 240 °C and 37 bar, 202 °C
and 27 bar and 108 °C and 15 bar in comparison with the corresponding impedance measurements can be
seen in Figure 37. The investigated cell showed current densities of 1000 mA cm− 2 at 240 °C, 640 mA cm− 2
at 202 °C and 85 mA cm− 2 at 108 °C at a cell voltage of 1.5 V (during decreasing the cell voltage). The
current densities at 1.75 V were 200 mA cm− 2 , 1440 mA cm− 2 and 1940 mA cm− 2 at temperatures of 108
°C, 202 °C and 240 °C, respectively. The results are not Rs - compensated. The impedance plots at a
polarization voltage of 1.4 V show that Rs improved from 164 mΩ cm2 to 128 mΩ cm2. ASRcell improved
more significantly from 508 mΩ cm2 to 361 mΩ cm2 with increasing temperature from 202 °C to 240 °C as
shown in the lower part of Figure 37. The impedance curve shows significant scattering at low frequencies;
a Kramers-Kronig test showed an error above 1 % (up to ± 4.5 %) at frequencies below 10 Hz with random
distribution around zero.
Figure 37 Cyclic voltammogram (upper figure) of a FobAEC (Cell 3a) with Ag- deposited anode and Inconel cathode at 240 °C and 37 bar (full line), 202 °C and 27 (dashed line) and 108 °C and 15 bar (dash-dot) along with the corresponding impedance analysis (lower figure) from data achieved at a polarization of 1.4 V.
Determination of the hydrogen and oxygen evolution overpotentials
The overpotential of the Inconel cathode has been studied with the use of the sample holder as shown in
Figure 24. The reference terminal of the potentiostat was connected to the Pt- wire in the center of the
FobAEC, see Figure 24. The Pt-wire acts as a reference electrode. The analysis of the hydrogen evolution
reaction, HER, has been performed in safety gas (9 % H2 in N2) with the results for the cyclic sweep
voltammogram shown in Figure 38. At a current density of 100 mA cm − 2 the overpotential was − 40 mV
and at 1000 mA cm − 2 it was − 131 mV, resulting in a Tafel-slope of 91 mV ⁄ dec. The overpotential
uncompensated for IR drop at a current density of 1 A cm − 2 was -0.28 V.
51
The IR- corrected results (ηH2) correspond to the difference of the cathode potential vs the equilibrium
potential of the reference electrode EeqC. The cathode potential vs the SHE (EC, SHE) is displayed on the
second ordinate of Figure 38.
Figure 38 Cyclic voltammogram for the hydrogen evolution overpotential of the Inconel cathode with a sweep rate of 20 mV/s at 42.1 bar and 248 °C. The full line shows the Rs- compensated applied cell voltage; the dashed line shows the graphical estimation of an Tafel line between 100 mA cm
− 2 and 1 A cm
− 2.
The oxygen overpotential has been measured in technical air (20 % O2, 80 % N2) for cells with and without
the Ag catalyst with the results shown in Figure 39. The obtained current density for low overpotentials
between 50 mV and 150 mV is higher for the cell without Ag, but this was not generally the case (compare
Figure 35). At higher overpotentials the situation changes and the current density of the Ag activated anode
is higher than that of the pure nickel foam anode, i.e. at a current density of 90 mA cm − 2 the overpotentials
are 250 mV and 220 mV, while at 900 mA cm − 2 they are 360 mV and 380 mV for the Ag activated anode
and the nickel anode, respectively. The Tafel slope for the not activated cell is 160 mV ⁄ decade, while the
Ag- activated anode improved to 130 mV ⁄ decade between 90 mA cm − 2 and 900 mA cm − 2. The curves
appear to be highly nonlinear.
The IR- corrected results (ηO2) correspond to the difference of the cathode potential vs the equilibrium
potential of the reference electrode EeqA. The anode potential vs the SHE (EA, SHE) is displayed on the second
ordinate of Figure 39.
52
Figure 39 IV-curve of a Ag loaded anode at 43.3 bar and 248 °C (full line) in comparison with an IV-curve of a pure nickel foam anode (dashed) at 42 bar 250 °C, both recorded with a sweep rate of 50 mV/s and compensated for Rs.
5.3.6. “Long term” galvanostatic test
The galvanostatic measurement on cell called cell 70 showed an activation period at the beginning of the
measurement. Ucell at tel = 0 h was 1.49 V, decreasing to 1.44 V at tel = 5 h, resulting in an activation rate of
10 mV ⁄ h for the first 5 h. With ongoing electrolysis time, cell voltage peaks (irregular oscillations)
occurred. The minimum and the maximum value of these peaks are shown in Figure 40 along with the
average cell voltage Ucell, mean for each 5 h lasting measurement period. It can be seen that a relatively
strong degradation takes place from tel = 5 h until tel = 45 h, with Ucell, mean starting from 1.45 V and
ending at 1.66 V. The resulting degradation rate in this period is 5.25 mV ⁄ h. From tel = 45 h until tel = 230
h the degradation rate decreased to 0.6 mV ⁄ h, ending at Ucell, mean = 1.77 V. Further on, the cell activated
again to Ucell, mean = 1.67 V at tel = 345 h with an activation rate of 0.9 mV ⁄ h.
Such an initial activation followed by a degradation process was also observed by Divisek et al. for an
electrolysis cell under galvanostatic conditions (400 mA cm − 2, 100 °C, 10 M KOH); corrosion problems were
indicated as the reason for degradation in their measurements. Although viewed in terms of Ucell, mean cell
70 seems to have activated during that last period, the observed cell voltage peak maximum and repeating
rate increased during this time. The galvanostatic impedance measurements which were performed at the
end of each 5 h interval led to too high cell voltages for the potentiostat and the measurement was
stopped after 350 h.
53
Figure 40 Development of the 5 h average of the applied cell voltage along the min and max values over the entire measurement period of 345 hours with a constant current density of 640 mA cm2, at 250 °C and 40 bar.
A fast Fourier transformation FFT of the raw data was performed in order to analyse if the peaks had one or
more specific repetition frequencies or if their periodicity varied randomly. It was concluded that the peaks
varied randomly. This may be interpreted as an effect of formation of probably big gas bubbles that stayed
on the hydrogen electrode for stochastically varying length of time.
5.3.7. Summary
Comparison of the results from the high temperature and pressure measurements to the results from the
measurements in the low temperature measurement setup displayed in Figure 34, gives an impression on
how significantly the current density of an alkaline electrolysis cell can be increased by increasing the
temperature. The analysed Raney-nickel foam electrode obtained a current density of 40 mA cm-2 at an
applied cell voltage of 1.75 V and a temperature of 100 °C. Cell 2 and 3a from Figure 35 reached current
densities of 1.57 A cm-2 and 2.0 A cm-2 at temperatures and pressures around 250 °C and 40 bar. The Raney-
nickel electrode reached a current density of 200 mA cm2 at a relatively high cell voltage of 2.24 V at 100 °C.
It should be stressed out that neither the Raney-nickel electrode nor the low temperature measurement
system has been optimized. Furthermore, it is clear that a significant improvement of the obtained current
densities is obtained in the high temperature measurements by comparing to literature data. Divisek et al.
(1988) reached a current density 400 mA cm2 at a cell voltage of 1.55 V and a temperature of 100 °C on a
Raney-nickel cell with both electrodes made of Raney-nickel. This is among the best reported in the
literature for alkaline electrolysis, but less than half of the performance obtained in this project at 250 C
and 40 bar.
54
Thus, it was demonstrated that increasing the operation temperature of an alkaline electrolysis cell can
reduce the losses significantly, thereby increasing the electrical efficiency and decreasing the hydrogen
production costs. Operation at elevated pressures can furthermore reduce the investment costs.
Measurement systems were designed and constructed to conduct electrochemical measurements at
temperatures up to 250 °C and a pressure of 95 bar.
5.3.8. List of publications of the project from DTU Energy Conversion
1. Frank Allebrod, “High temperature and pressure alkaline electrolysis”, PhD thesis, DTU Energy Conversion, November 2012. The thesis will be available on the DTU web after defense and acceptance, which is scheduled to take place in January 2013. A printed copy is available for perusal at the main supervisor, Mogens B. Mogensen, DTU Energy Conversion.
2. Frank Allebrod, Christodoulos Chatzichristodoulou, Pia Lolk Mollerup, Mogens Bjerg Mogensen, “Electrical conductivity measurements of aqueous and immobilized potassium hydroxide”, International Journal of Hydrogen Energy, 37 (2012) 16505 – 16514.
3. Frank Allebrod, Christodoulos Chatzichristodoulou, Mogens B. Mogensen, “Alkaline Electrolysis Cell at High Temperature and Pressure of 250 °C and 42 bar”, submitted to Journal of Power Sources 2012, accepted with conditions.
4. Frank Allebrod, Christodoulos Chatzichristodoulou, Pia Lolk Mollerup, Mogens Mogensen, “High Performance Reversible Electrochemical Cell for H2O Electrolysis or conversion of CO2 and H2O to Fuel”, Patent Application nr. EP12164019.7, Date of filing: 13/04/2012.
5.4. Results on improved electrodes
Work package 1 has covered the development of three different electrode tracks. The aim of this WP has
been to cover the two first objectives of the project, to increase electrode efficiency to more than 88% at a
current density of 200 mA /cm2 as well as increase the operation temperature to more than 100 degree
Celsius to make the cooling energy more valuable. The findings from the three electrode technology tracks
are summarized in the following.
From the beginning only electroplating and the atmospheric plasma spray technology have been
considered as competing technologies in a pre-commercial scale. The HTP track is still on a research stage
and further development is needed before the electrode is implemented in a cell that will be an option for
commercial use.
5.4.1. Electroplating
During the presented Ph.D. study, at DTU-ME a new type of electrodes, suitable for industrial alkaline water
electrolysers, has been developed. The electrodes are produced by magnetron sputtering a thin film of
aluminium onto a nickel or nickel plated substrate. Hereafter, the aluminium is interdiffused into the nickel
surface, by heating, and several intermetallic phases between Al and Ni are formed. Lastly, aluminium is
selectively leached from one of the intermetallic phases on the surface and high porous nickel electrode is
formed.
55
The structure of the developed electrodes has been characterised via high resolution scanning electron
microscopes, revealing large actual surface area with pores down to 20 nm.
Consistent electrochemical measurements show that the developed electrodes need 450 mV less voltage
for decomposing water into hydrogen and oxygen compared to smooth nickel electrodes, in 1 M KOH at 25
°C and 1 atm.
Durability test in an electrolysis stack for 9,000 hours indicated no serious mitigation in the electrode
efficiency during the test period. The efficiency of the electrolysis stack was measured to be around 78-79
% compared to the higher heating value (HHV) at 65 °C. It should be mentioned that the electrolysis stack
was of an old version and some efficiency loss was caused by the stack itself and not by the electrodes. The
same stack was tested at H2College at a temperature of 74.4 °C. Here an efficiency of 86.5% was obtained.
5.4.2. Atmospheric Plasma Spray
Quite early in the project conducted by FORCE Technology, it was shown that highly efficient cathode
coatings meeting the goal of 88 % efficiency could be produced by Atmospheric Plasma Spraying (APS). At
the same time, FORCE Technology established a test concept for characterising the electrode coatings that
verified APS coatings as fully comparable or better than those described in the literature. Both efforts
placed the project right on track from the beginning.
Further developments of the APS coatings led to additional improvements in stability of substrate,
adhesion to substrate, spraying parameters and optimisation of coating thickness.
Larger electrodes in many shapes were provided for the pilot test stacks with promising results, though the
reproducibility was not necessarily evident. Near the end of year two, only marginal improvements were
made on the NiAl coatings for the cathode. At this stage, the cathode side was considered nearly or fully
developed as evidenced by the impressive efficiency reported for the pilot cells (97 %). The subsequent
development of coatings for the anode side did not provide the desired results, but gained important
insight about possible future process routes.
The developed NiAl electrodes produced by APS have made major steps forward to be suitable for large-
scale production. However further long-term test for verifying the durability and life time still have to be
conducted. Furthermore price reductions are still needed to make the technology competitive. This
requirement might be met by going for a large scale production. However, this will require investments and
establishment of a company outside FT auspices.
5.4.3. HTP electrodes
In the Ph.D. study at DTU-EC, comparison of the results from a high temperature and pressure
measurements with the results from the measurements in a low temperature setup have shown how
significantly the current density of an alkaline electrolysis cell can be increased by increasing the
temperature. Analyses have shown Raney-nickel foam electrode can obtain a current density of 40 mA cm-2
at an applied cell voltage of 1.75 V and a temperature of 100 °C. Further the Raney-nickel electrode has
56
reached a current density of 200 mA cm2 at a relatively high cell voltage of 2.24 V at 100 °C. In comparison
current densities of 1.57 A cm-2 and 2.0 A cm-2 have been reached when temperatures around 250 °C and
pressures at 40 bar have been applied to the HTP electrode. It should be stressed that neither the Raney-
nickel electrode nor the low temperature measurement system has been optimized. Furthermore, from
literature data it is clear that a significant improvement in current densities is obtained in a high
temperature scenario. Divisek et al. (1988) reached a current density 400 mA cm2 at a cell voltage of 1.55 V
and a temperature of 100 °C on a Raney-nickel cell with both electrodes made of Raney-nickel. This is
among the best reported in the literature for alkaline electrolysis, but less than half of the performance
obtained in this project at 250 C and 40 bar.
Thus, the project has demonstrated that increasing the operation temperature of an alkaline electrolysis
cell can reduce the losses significantly, thereby increasing the electrical efficiency and decreasing the
hydrogen production costs. Operation at elevated pressures can furthermore reduce the investment costs.
However as these results are obtained in a research lab on a cell measuring 1 cm2, further developments
are required to demonstrate the findings in a full scale cell, stack and system.
5.4.4. Conclusion
The results obtained in this work package show that the electroplating track on the short term gives the
most promising performance when combining durability, lifetime, efficiency and price.
For the APS method higher cell efficiencies has been reached, however long term test on durability is still
needed as well as further reduction in price is required.
Finally, the HTP track has shown that elevated process temperature causes remarkably higher current
density entailing higher cell efficiency.
57
6. Higher operation pressure (WP 3)
The goal set for the second generation electrolyser system, has been to generate 30 bar pressure in the cell
stack. Operating at elevated pressure have several advantages, but also proves to have some challenges. In
this work package solutions for overcoming the challenges has been developed.
When water is split by electrolysis, hydrogen is naturally generated at a very high pressure. If the generated
hydrogen is contained, it is possible to generate compressed hydrogen directly from the cell stack, without
spending additional electricity. This is a great advantage, as the hydrogen can be pressurized up to 30 bar,
without using a conventional compressor, which is both energy consuming, and costly. While some
applications will need further compression for storing the hydrogen, it will be possible for many
applications to work completely without a compressor, greatly simplifying and improving the overall
system.
Another great advantage of high pressure electrolysis is the reduced size of tubing and components.
Compared to hydrogen at atmospheric pressure the reduced volume of the compressed hydrogen
provide a possibility to increase the capacity or reduce the size of the components by a factor equal to the
pressure ratio, and still retain the same flow rates and roughly the same pressure drops. This enables the
electrolyser system to be designed with a significantly smaller footprint than equivalent atmospheric
systems. The reduction of components size provides a possibility to modularize the electrolyser concept, as
well as reducing the price of the individual components.
One of the drawbacks of high pressure hydrogen electrolysis is increasing demands on materials and stack
design, as the electrolysis stack will need to be manufactured primarily from a nonconductive material
which is chemically stable against hot KOH, H2 and O2, and still satisfy the mechanical requirements for
containing electrolysis under 30 bar.
To accommodate the higher pressure, the stack has been designed as a cylindrical pressure vessel, with
each cell having a cell “wall” sufficiently thick, to resist the high pressure and sealed with O-rings for perfect
sealing at high pressures.
For a more detailed description of the stack design, see chapter 7.
Another challenge for high pressure electrolysis has been equalisation of the H2 and O2 pressure. As with
electrolysis under atmospheric pressure, it is important to control the pressure on the H2 and O2 side of the
electrolyser, to ensure that no mixing of gasses can occur. At atmospheric pressure the system is controlled
passively with gravity ensuring the system stays in balance, hence no pressure difference can drive gas to
cross the diaphragm or cause emptying
or flooding of the separator or scrubber
vessels.
At elevated pressures however, it
becomes increasingly difficult to control
the pressures within milli-bars, which is
necessary for maintaining the water
58
levels of the separator and scrubber vessels in equilibrium.
To solve this problem, a special equilibrium valve has been developed to mechanically control that the
pressure of the H2 at all times equals the O2 side . The valve was designed in such a way, that even a slight
difference in pressure between the sides, causes the under-pressured side to close. Thus pressure will be
build up until it is equal to the over-pressured side. This regulation is constant and gradual, and will
maintain the pressure difference between the sides within milli-bars.
59
7. Improved electrolysis stack architecture (WP 4)
The primary goal of WP4 has been to develop an electrolyser stack concept suitable for mass production,
while still solving the technical challenges when designing a high pressure alkaline electrolyser stack.
The challenges for construction of a high pressure alkaline electrolyser stack are mechanical strength,
electrolyte and gas flow, stray currents and galvanic corrosion.
The first design decision after deciding to develop a high pressure electrolysis stack, was to decide on a zero
gap or non-zero gap design. The advantages of a zero gab design, is the improved efficiency of the
electrolyser, because of a reduced distance between electrodes. However a zero gap design is also very
costly, primarily because the additional materials and production costs for zero gap electrodes.
When operating an electrolyser at a higher pressure, it is possible to reduce the distance between
electrodes, because of smaller volumes of gas being released. This factor is very important for the design of
a high pressure alkaline electrolyser, because it enables the use of the much simpler non-zero gap design.
For this design to be viable however, it is necessary to keep the active area of each electrode sufficiently
small, to keep the volumetric ratio of gas to electrolyte low and to maintain a high conductivity in the
electrolyte, even at the top of the electrolysis cells where gas bubbles will naturally accumulate.
It was from these considerations, the concept of a “low gap”, low diameter, high pressure and high cell
number electrolyser stack was born.
For the scope of mass production and cost reduction, this design philosophy is also ideal, because it allows
the use of only very few different parts, with a relatively small size, in large quantities.
Locked in to this design philosophy, it was now necessary to consider the next design constraint on the
electrolyser stack: The mechanical strength.
To achieve a sufficiently high mechanical strength of the electrolyser stack, several approaches was
researched, and it was settled on using a high strength polymer for the bulk cell material. The cells have
therefore been designed with a sufficient wall thickness to be tested up to 40bar.
The great advantage of using a polymer material is the possibility for mass production by injection
moulding. However, great care has been taken when selecting material for injection moulding of cell
frames, as reinforcements and flow fronts during moulding are very important for the overall strength of
the material.
Related to the material strength of the cell stack, is also the thermal expansion of the cell frames with
increasing temperature. Because of the uneven expansion of the cell frames and the tie rods of the
electrolyser, it is necessary to use disc springs to allow thermal expansion of the electrolysis stack.
In an electrochemical system an electric circuit is formed by making a bridge through an electrolyte. For the
case of alkaline water electrolysis, the circuit is formed by transporting OH- from the cathode to the anode,
releasing H2 and O2 respectively.
60
This mechanic however, results in some technical difficulty when it comes to designing an alkaline
electrolysis stack, because an electrolyser of the bipolar design has shared manifolds. Each electrode can
here see all other electrodes through the ports and manifolds of the system, which result in unintended
currents through ports and manifolds.
These currents are commonly named either stray currents, shunt currents or leak currents.
Dimensioning for stray currents was part of the initial considerations, as a design was proposed with long
ports with a low cross sectional area. However, during experiments of the stack, it was understood that the
stray currents was significantly increased with a higher cell number than what was expected from initial
calculations, even though steps was taken to reduce the effect during cell design.
The challenges of limiting stray currents was more comprehensive than initially expected and the result is
that alternative methods for reducing stray currents in high cell number stacks has been designed for
testing in future generations of the high pressure alkaline electrolyser stack. And a more detailed
understanding of the stray current phenomena has been used to develop more accurate computer models.
61
8. Corrosion resistant materials (WP 5)
If hydrogen is to be produced at increased temperature, pressure and stack voltage, it requires that the
electrolyser can withstand the aggressive alkaline media under those severe conditions. A design life of
more than 10 years is a challenge in terms of materials. The corrosion mechanisms of metallic materials
have previously been studied under similar conditions, and literature data and experience were collected to
identify critical risks of corrosion and stress corrosion cracking. Corrosion may be a problem for the
electrolyser components both in terms of poisoning the electrodes (iron contamination) and structural risk.
This reduces the options for metallic materials to primarily nickel-based alloys, if the cell should run at the
defined conditions (30 % KOH, 100 °C, 20 bar). If austenitic stainless steels are to be used, the conditions
must be less severe. The data were applied in the design of experimental cells.
Summary of corrosion tests in autoclave at 120 °C in 25 % KOH for 14 days.
Material type Pass/Fail Comments Application
PPS Fail Glass fibres dissolved, cracks and dimensional and weight change
Cell frame
PA Fail Decomposed Cell frame
PEI Fail Decomposed Cell frame
PPS Fail Blisters Cell frame
PFA Pass Slight discoloration Cell frame
PEEK Pass No sign of degradation Cell frame
PEEK+CA Pass No sign of degradation Cell frame
PESU Pass Small blister at cut face Cell frame
PS Pass No sign of degradation* Diaphragm
PS Pass Slight weight gain Diaphragm
PS + Zr oxide (90C)
Fail Decomposed* Diaphragm
PS + Zr oxide (120C)
Pass No sign of degradation* Diaphragm
PS + Zr oxide (500)
Pass No sign of degradation Diaphragm
PS + Zr oxide (550)
Pass No sign of degradation Diaphragm
Table 5 Summary of corrosion tests in autoclave at 120 °C in 25 % KOH for 14 days. *7 days test only as autoclave cracked by stress corrosion cracking in test.
The initial literature survey and the first tests at H2College also indicated that the chemical resistance of
polymers presented a greater challenge than anticipated, and that test data from commercial suppliers
were insufficient to model the conditions in the electrolyser. The alkali resistant polymers (e.g. Teflon) are
62
costly and the search for cheaper alternatives turned into a major aim. The progress of the project
therefore required corrosion testing and examination of polymers to find alternative durable membrane
and gasket materials. A number of different tests were run under accelerated conditions and the
degradation mechanism was examined. The results of standardised tests are summarised in Table 5.
Consequently, the project has involved corrosion testing and examination of polymers to greater extent than
originally planned at the detriment of long-term testing of electrodes (WP 1.8), which has been limited.
Inspiration for the future materials selection in electrolysers is now available by the combined experience
from the experimental work and literature.
63
9. Modular design for easy up scaling (WP 6)
One of the strong visions for the second generation electrolyser concept was to develop a modular system
with several standardized components, to make it possible to deliver a wide range of specification and
features to the customer depending on the application.
The use of a modular system is also aimed at making service faster and easier, by making it possible to
replace defect modules directly on site to decrease the downtime. And also centralize the repair of defect
modules to the company workshop, greatly reducing
the need for mobile tools and technicians.
A great effort has been put into analysing the
electrolysis system to determine how the different
components could be split onto modules. The end
result is that following modules have been designed.
Electrolyser module
Deoxer Module
Dryer Module
Water Treatment Module
Power supply and Control unit
Rack mount
The Electrolyser, Deoxer and Dryer module have been
designed and manufactured from scratch, where the
power supply and control unit, water treatment module
and rack mount have been manufactured from existing
and standardized components.
9.1. Electrolyser module
The purpose of this module is to convert water into hydrogen by applying a DC current to a water based
electrolyte. The splitting of the water occurs in the stack component, where electrolyte (lye, KOH) is
continuously fed into a channel in the bottom. The stack consists of a number of cells, where the H2 and O2
production occurs. The cells are all connected by channels, where a common H2 and O2 outlet is connected
to a H2 and an O2 separator vessel. During electrolysis, the electrolyte will be filled with small gas baubles.
These bubbles will reduce the density of the electrolyte, and thereby cause a circulation, since the density
of the lye in the return flow is higher. The electrolyte/gas mixture is thus pumped into the separator
vessels.
64
In the separator vessels, the gases are separated from the electrolyte, and afterwards the hydrogen and
oxygen is scrubbed with demineralized water from the water supply to remove residual amounts of lye
from the gas. After separation, the gasses exits the module.
The control concept for the electrolysis module is based on an equalization valve that only allows flow of
both gasses, as long as the hydrogen and oxygen have equal pressure. A pressure drop in one of the gasses
will result in the valve blocking the flow, until pressures become equal again. This will ensure that the
gasses can’t pass through the diaphragm
and mix with each other.
A nitrogen purge channel is connected to
both separator vessels in order to purge
the system during startup and when
service is needed.
The water consumed by the electrolysis
process is refilled into the scrubber
compartments of the separator vessels.
There will thus be a constant refill with
demineralized water from this
compartment into the separators, which
ensures the water supply for the electrolysis.
9.2. Deoxer Module
The primary function of this module is to remove oxygen from the hydrogen by catalytically burning the
oxygen and secondarily drain water from the system.
The wet hydrogen is initially cooled and excess water is
drained, to avoid liquid water from causing corrosion
within the deoxer and inhibit the catalytic process. The
hydrogen is then preheated by the ambient
temperature in the rack before entering the catalytic
burner (Deoxer),
When the condensed water is taped a small amount of
hydrogen is released to the atmosphere at the same
time.
The deox module is also responsible for controlling the
purity of the oxygen and hydrogen for safety reasons,
to avoid generation of explosive gas mixtures.
65
9.3. Dryer Module
Drying is accomplished by leading the product gas through a zeolit that adsorbs water vapour, drying the
gas down to a dew point of -80C.
The zeolites are slowly saturated with water, and will therefore need regeneration.
Regeneration is done by utilizing a dual drying system, where one dryer dries the product gas, while the
other dryer regenerates.
To regenerate a dryer, some of the dry product gas is led back through the dryer at low pressure, to carry
away desorbing water from the zeolite. (the zeolite cannot contain much water at low pressure, and thus
the water is desorbed.)
The wet regenerated gas is then
vented to the atmosphere
resulting in loss of overall system
efficiency. The amount of gas
needed for regenerating the dryer,
is equal to the pressure ratio
between the absorption pressure
and the desorption pressure.
Hence the higher pressure in the
cell stack the less gas is needed for
regeneration.
This loss of overall efficiency is not
desired for systems that do not
depend on hydrogen with very low
water content. The dryer is
therefore an optional module for
these applications.
9.4. Water Treatment Module
This module contains the water treatment for the system.
Because water is split in the electrolyser, water will continuously need to be fed to the electrolyzer module.
Which means any impurities in the water will increase in concentration during electrolysis. Some
impurities, especially Cl have very damaging effects on the electrodes and it is therefore important to purify
the water before it enters the system.
66
The water is led through a mix-bed ion exchanger in order to achieve a conductivity between 0,1-1
microSiemens. Depending on the water supply at the site, it can be necessary to implement a reverse
osmosis unit for pre-treating the water before leading it into the mix-bed ion exchanger.
To feed the water into the pressurized electrolyser, the water is pressurized with a pump and led into a
buffer vessel.
9.5. Power supply and Control unit
The power supply was built as simple as possible and consist of a number of purchased AC/DC converters.
The control unit is based on standard PLC controls.
This solution was selected because many good products are already available, and the area of power supply
has not been the focus of this project.
The power supply and control system was for the H2 College electrolyser placed in a separate compartment
for safety reasons.
9.6. Rack Mounted Electrolyser
The goal was to design the electrolysis system so it would fit into standardized racks, to standardize the
footprint measurements, and ease installation requirements.
This goal was achieved, so the only requirement currently, is available floor space, electricity, water, and
pipes for hydrogen, oxygen and ventilation.
On top of the rack is mounted a blower, to apply constant suction to the electrolyser during operation, so
to ensure any potential leak cannot build up to create an explosive atmosphere.
The end goal to be able to fit up to 8 Nm3 electrolyser capacity into one rack (though currently the power
supply and controls have been placed into a separate rack) is firmly believed possible, by fitting the controls
and power supply into the bottom of the electrolyser rack, once the development leaves the prototype
stage.
67
10. Demonstration (WP7)
Image 1: Electrolyser building at H2 college
10.1. Building and Installation
For containing the electrolyser and showcase hydrogen technology, a house was built in conjunction with
the 66 H2 college houses. The electrolyser system was installed in winter 2010, with the entire project
finalized early spring 2010. Once the ground thawed and it was possible to dig down the connecting pipe
from the electrolyser building to the H2 college houses the plant was ready to deliver hydrogen to the fuel
cells.
The installation of the electrolyser system was as simple as the modules had already been assembled from
the workshop, so installation was limited to fixing the modules in the rack, connecting power, water, gas
pipes and ventilation, and filling KOH to the electrolyser module.
Because of the prototype nature of the electrolyser, the controls for the system had not been finalized
before the installation, so for the case of H2 college project, the commissioning of the plant was performed
68
on site. For future plants however, the commissioning will take place in the workshop, with only minimal
adjustments if any, needing to be done one the site.
Image 2: Electrolyser system and storage tank
10.2. Measurement Results
10.2.1. Stack efficiency
A number of measurements were performed on the system installed in H2College, showing a stack
efficiency of 86.5% at a current density of 177mA/cm2 and a temperature of 74.4 ˚C at the hydrogen outlet.
10.2.2. System Efficiency
There have not been any measurements of the overall system efficiency
10.2.3. Total production hours
After installation the electrolyser system is estimated to have run roughly 2000h
10.2.4. Gas purity
The purity of the gas was testes before entering the deoxer to 0,3% oxygen in hydrogen, which is more than
10 times lower than the critical 4%.
69
10.2.5. Cooling
The stack was not actively cooled, as passive cooling of the stack was sufficient to keep it at low
temperatures. Conversely, it was required to insulate the system to achieve desired operating
temperatures.
The cooling of product gas in the deox module have worked as intended, as no condensation of water has
been detected after the deox module. For this reason it is assumed that it is not necessary to use a dryer
module in applications where no additional compression is required.
10.3. Demonstration
The demonstration of the second generation alkaline electrolyser is considered a success for several
reasons. While not all technical goals were reached, the research and development in the project have
resulted in some very good solutions for most of the challenges in high pressure alkaline electrolysis.
The concept of a rack mounted electrolyser was proven viable, and all critical aspects of the modularisation
have been covered.
What H2 College demonstrates is therefore more than the development of a small size electrolyser, but it is
a leap from developing electrolyser systems as a plant, to developing electrolyser systems as a product
suitable for mass production.
70
11. List of publications
The findings of the project have been communicated in different ways. The following list embraces
publications as articles in international journals as well as presentations and posters on national and
international conferences.
‘Modelling electrolyte conductivity in a water electrolyzer cell’, , Michael Caspersen(a), Technical University
of Denmark, Julius Bier Kirkegaard(b), Niels Bohr Institute, University of Copenhagen,
International Journal of Hydrogen Energy - Volume 37, Issue 9, May 2012, Pages 7436–7441
‘Development of durable and efficient electrodes for large-scale alkaline water electrolysis‘, Cecilía
Kjartansdóttir, [email protected],Technical University of Denmark, Lars Pleth Nielsen, [email protected],
Danish Technological Institute, Per Møller, [email protected], Technical University of Denmark,
Submitted to International Journal of Hydrogen Energy January 31 2013.
‘2nd Generation Alkaline Electrolysis for Hydrogen Production‘,Cecilía Kjartansdóttir,
[email protected],Technical University of Denmark, Lars Pleth Nielsen, [email protected], Danish
Technological Institute, Per Møller, [email protected], Technical University of Denmark,
Presented at the Metallurgisymposium, Roskilde on June 9 2011.
‘Electrodes with good durability for alkaline water electrolysis’, Cecilía Kjartansdóttir,
[email protected],Technical University of Denmark, Lars Pleth Nielsen, [email protected], Danish
Technological Institute, Per Møller, [email protected], Technical University of Denmark,
Poster presentation at World Hydrogen Energy Conference 2012, Toronto Canada June 2012
‘Electrolysis plants heat balance’, Lars Yde, [email protected]
Presentation at SYMPOSIUM on Water electrolysis and hydrogen as part of the future Renewable Energy
System, DTU, Risø Campus
‘Atmospheric plasma spraying of electrodes for conversion of electricity to hydrogen by high temperature alkaline electrolysis of water’ Peter Tommy Nielsen, [email protected], Troels Mathiesen, Jens Klæstrup Kristensen, Lisbeth Hilbert, FORCE Technology Denmark, Lars Yde, HIRC Presentation and proceedings at International Thermal Spray Conference 2011, Hamburg September 2011
’Electrochemical characterisation of atmospheric plasma spray coatings for water electrolysis’ Troels Mathiesen, [email protected], Lisbeth Hilbert, Peter Tommy Nielsen, FORCE Technology Denmark
Poster presentation at Symposium: Water Eletrolysis and hydrogen as a part of the future Renewable
Energy System, Copenhagen May 2012. Also presented at DEF Electrochemical Science & Technology 2012,
October 2012
‘High Temperature and pressure alkaline electrolysis’, Frank Allebrod, [email protected],Technical University of
Denmark, Sune D Ebbesen, [email protected], Technical University of Denmark, Johan Hjelm, [email protected],
Technical University of Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, Oral
presentation at the Conference on Electrochemical Science and Technology, Roskilde, Denmark, October 1st
2009.
71
‘Determining the Conductivity of Aqueous Electrolytes using the van der Pauw Method‘ Frank Allebrod,
[email protected],Technical University of Denmark, Pia Lolk Mollerup, [email protected], Technical University of
Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, Poster presentation at the
HyFC Academy summer school, Roskilde, Denmark, August 22nd
– 26th 2010.
‘Conductivity Measurements of Aqueous and Immobilized Potassium Hydroxide ‘ Frank Allebrod,
[email protected],Technical University of Denmark, Pia Lolk Mollerup, [email protected], Technical University of
Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, Poster presentation at the
HyFC academy research school, workshop & annual meeting, Roskilde, Denmark, April 11th - 15th 2011.
‘Electrolysis for Integration of Renewable Electricity and Routes towards Sustainable Fuels’, Mogens
Mogensen, [email protected], Technical University of Denmark, Frank Allebrod, [email protected],Technical
University of Denmark, Jacob R. Bowen, [email protected],Technical University of Denmark, Christodoulos
Chatzichristodoulou, [email protected],Technical University of Denmark, Ming Chen, [email protected],Technical
University of Denmark, Sune D Ebbesen, [email protected], Technical University of Denmark, Christopher R.
Graves, [email protected],Technical University of Denmark, Jonathan Hallinder, [email protected],Technical
University of Denmark, Anne Hauch, [email protected],Technical University of Denmark, Peter Vang
Hendriksen, [email protected],Technical University of Denmark, Peter Holtappels, [email protected],Technical
University of Denmark, Jens Valdemar Thorvald Høgh, [email protected],Technical University of Denmark, S.H.
Jensen, [email protected],Technical University of Denmark, Alberto Lapina, [email protected],Technical University of
Denmark, Pia Lolk Mollerup, [email protected],Technical University of Denmark, Xiufu Sun, [email protected],
Technical University of Denmark, 10th International Workshop on Large-Scale Integration of Wind Power
into Power Systems as well as on Transmission Networks for Offshore Wind Farms, 2011, isbn: 978-3-
9813870-3-2.
‘Electrical conductivity measurements of aqueous and immobilized potassium hydroxide’, Frank Allebrod,
[email protected],Technical University of Denmark, [email protected], Christodoulos Chatzichristodoulou,
[email protected],Technical University of Denmark, Pia Lolk Mollerup, [email protected], Technical University of
Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, Proceedings of the
International Conference on Hydrogen Production, ICH2P-11, Paper No 181ELE , June 19-22, 2011,
Thessaloniki, Greece.
‘Electrical conductivity measurements of aqueous and immobilized potassium hydroxide’, Frank Allebrod,
[email protected],Technical University of Denmark, Christodoulos Chatzichristodoulou, [email protected],Technical
University of Denmark, Pia Lolk Mollerup, [email protected], Technical University of Denmark, Mogens
Mogensen, [email protected], Technical University of Denmark, International Journal of Hydrogen Energy,
Volume 37, Issue 21, Pages 16505 – 16514, November 2012.
‘High Performance Reversible Electrochemical Cell for H2O Electrolysis or Conversion of CO2 and H2O to Fuel’,
Frank Allebrod, [email protected], Technical University of Denmark, Christodoulos Chatzichristodoulou,
[email protected],Technical University of Denmark, Pia Lolk Mollerup, [email protected], Technical University of
Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, patent Pending No.
12164019.7-2119, April 2012.
72
‘FobAEC – Foam based Alkaline Electrochemical Cell for High Temperature and Pressure Alkaline Electrolysis’
Frank Allebrod, [email protected],Technical University of Denmark, Christodoulos Chatzichristodoulou,
[email protected],Technical University of Denmark, Mogens Mogensen, [email protected], Technical University of
Denmark, Poster presentation at the Water electrolysis and hydrogen as part of the future Renewable
Energy System symposium, Copenhagen, Denmark, May 10th- May 11th 2012.
‘High temperature and pressure alkaline electrolysis cell’, Frank Allebrod, [email protected],Technical University
of Denmark, Christodoulos Chatzichristodoulou, [email protected],Technical University of Denmark, Mogens
Mogensen, [email protected], Technical University of Denmark, Oral presentation at the World Hydrogen
Energy Conference 2012, Toronto, Canada, June 3rd - June 7th 2012.
‘Alkaline Electrolysis Cell at High Temperature and Pressure of 250 °C and 42 bar’, Frank Allebrod,
[email protected],Technical University of Denmark, Christodoulos Chatzichristodoulou, [email protected],Technical
University of Denmark, Mogens Mogensen, [email protected], Technical University of Denmark, Journal of
Power Sources, Volume 229, pages 22-31, May 2013.
‘High Temperature and Pressure Alkaline Electrolysis’, Frank Allebrod, [email protected],Technical University of
Denmark, PhD Thesis, Technical University of Denmark, Risø Campus, Roskilde, Denmark, submitted
November 14th 2012 and defended February 28th 2013.